Scoping Review on Deep Learning Model for Classification and Prediction of Diabetes

Scoping Review on Deep Learning Model for Classification and Prediction of Diabetes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Systematic Review Scoping Review on Deep Learning Model for Classification and Prediction of Diabetes Adedayo Emmanuel OJO, Rotimi Felix AFOLABI, William BALOGUN This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9081794/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Diabetes mellitus is a growing global health challenge, particularly in low- and middle-income countries, where delayed diagnosis and inadequate treatment contribute significantly to complications and mortality. Artificial intelligence, particularly deep learning (DL), has emerged as a promising tool for improving diabetes classification and outcome prediction. Objective This scoping review aimed to map existing evidence on the development and application of one-stage and two-stage deep learning models for the classification and prediction of Type 1 and Type 2 diabetes. Methods A scoping review was conducted using PubMed and Google Scholar databases, guided by the Population-Concept-Context (PCC) framework and PRISMA-ScR methodology. Studies were included if they applied deep learning models to the classification and/or prediction of diabetes. Data extraction was performed using a structured spreadsheet capturing model type, dataset, features, and performance metrics. Results Out of 750 identified studies, 50 met the inclusion criteria. Convolutional neural network-based architectures were the most common (16; 38%), followed by recurrent neural networks and hybrid models. The majority of studies (43; 86%) used a one-stage deep learning approach integrating classification and prediction into a single step. Only 7 studies (14%) employed a two-stage framework, and none were conducted in the African context. Common datasets included the Pima Indian dataset and the UCI Machine Learning Repository, with limited use of local or clinical datasets. Conclusion Deep learning models demonstrate strong potential for improving diabetes diagnosis and prediction. However, the dominance of one-stage models and the lack of African-based studies highlight critical methodological and geographical gaps. Future research should explore two-stage models tailored to local datasets to enhance clinical relevance and promote global equity in AI-based diabetes care. Endocrinology & Metabolism Epidemiology Artificial Intelligence and Machine Learning Diabetes Deep learning Two-stage model Figures Figure 1 Figure 2 Figure 3 Introduction Diabetes mellitus is a chronic metabolic disorder characterized by persistent hyperglycemia, resulting from defects in insulin secretion, insulin action, or both ( 18 ). In 2021, it accounted for more than 2 million deaths globally and is a major cause of complications such as blindness, kidney failure, heart attacks, stroke, and diabetic foot ulcer ( 11 , 18 ). As of 2024, global prevalence is 11.1% (588.7 million people), and about 45.8% remain undiagnosed ( 5 , 10 ). In Nigeria, prevalence among adults is estimated at 3.7% ( 5 ). Type 2 diabetes accounts for more than 90% of all cases ( 8 ). Traditional diagnostics such as HbA1c and fasting glucose often fail to detect ( 15 , 17 ). Artificial intelligence (AI) has gained increasing attention in medical diagnostics due to its ability to analyze complex and high-dimensional clinical data beyond the capacity of traditional diagnostic approaches ( 9 ). Conventional methods for diagnosing chronic noncommunicable diseases, such as diabetes mellitus, often depend on fixed clinical thresholds and expert interpretation, which may limit early detection and consistency across populations ( 12 ). AI-based techniques, particularly machine learning and deep learning models, offer a complementary approach by learning hidden patterns from biomedical data and integrating multiple risk factors simultaneously. Moreover, recent advances seek to improve clinical relevance by incorporating medical knowledge into AI models, addressing concerns related to interpretability and trust. As a result, AI presents a promising framework for supporting early diagnosis, risk assessment, and decision-making in the management of diabetes mellitus. ( 9 , 12 , 14 ). Deep learning leverages multilayer neural networks to discover patterns in structured health data ( 1 , 7 ). Previous studies have suggested the use of one-stage DL models, performing prediction or classification in a single step. Two-stage models classify diabetes and then predict outcomes; they are easier to interpret ( 6 ). The need for sequential, clinically interpretable models has led to growing interest in Two-stage DL architectures, where the first stage performs classification (Type 1 vs. Type 2) and the second stage predicts complications or disease progression. This scoping review aimed to explore and map the existing literature on DL models for diabetes classification and prediction. In doing so, it seeks to identify research gaps and provide direction for future studies, particularly those addressing local contexts in underserved regions. Methodology Study Design A scoping review was conducted using the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines to map existing evidence on the use of deep learning models for diabetes classification and prediction. Research Question What deep learning approaches have been used to classify Type 1 and Type 2 diabetes, and how are they applied and evaluated in existing literature Search Strategy A systematic search was performed in PubMed and Google Scholar databases using Boolean search operators: "deep learning" AND "model" AND "diabetes". The search included studies applying deep learning to the classification and/or prediction of diabetes, without restriction on publication date, language, country, or population ethnicity. The process of study identification, screening, eligibility assessment, and final inclusion of studies is illustrated in Fig. 1 . Eligibility Criteria Studies were eligible for inclusion if they involved individuals diagnosed with Type 1 or Type 2 diabetes mellitus, applied deep learning or hybrid artificial intelligence models for the classification and/or prediction of diabetes, and were conducted within healthcare or diagnostic contexts globally. Only studies published between January 2017 and August 2025 were considered. Studies were excluded if they focused exclusively on traditional machine learning methods without deep learning components, involved non-human subjects, or addressed disease contexts unrelated to diabetes mellitus. In addition, some studies demonstrating two-stage modeling frameworks in biomedical or predictive analytics contexts were reviewed to illustrate methodological approaches relevant to diabetes prediction. Study Selection and Screening Initial records were screened for duplicates, followed by title and abstract screening. Eligible full-text articles were assessed for inclusion. Screening and extraction were performed by two independent reviewers. Data Extraction Data extraction was conducted using a structured Excel spreadsheet to systematically chart key study characteristics. Extracted information included the author(s), year of publication, country of study, the dataset used, the type of deep learning model applied, whether a one-stage or two-stage classification approach was employed, and the reported performance metrics. Results A total of 750 articles were identified. After removing 97 duplicates and screening titles and abstracts, 248 articles were retrieved for full-text review. Of these, 198 articles were excluded based on the eligibility criteria. A total of 50 studies met the inclusion criteria and were included in this scoping review. The studies varied in terms of data sources, deep learning architectures, and evaluation metrics used for diabetes classification and prediction. The characteristics of the included studies are summarized in Table 1 . Table 1 Characteristics of studies included in the scoping review. S/N Title Authors Year Country Dataset used Model/Method Key Focus Performance Metrics 1 Fully Automated Abdominal CT Biomarkers for Type Diabetes Using Deep Learning Tallam et al. 2022 United States CT imaging and clinical data 3D U-Net + Logistic Regression CT biomarkers for diabetes prediction Dice Score 0.69, AUC (Not specified) 2 A Deep Learning Model for Screening Type 2 Diabetes from Retinal Photographs Yun et al. 2021 United States Retinal images (UK Biobank) ResNet18 fine-tuned Retinal-based diabetes prediction AUC 73% (algorithm), 84% (with traditional risk factors) 3 A Deep Learning Model Incorporating Knowledge Representation Vectors for Diabetes Prediction Xu et al. 2022 China Clinical examination data THSAC model (self-attention + CNN) Knowledge-based deep learning for diabetes Accuracy (Not specified), F1-score high 4 A novel Hybrid Deep Learning Model for Early-Stage Diabetes Risk Prediction Bülbül 2024 Turkey UCI Diabetes dataset Hybrid deep learning model (GA + Autoencoder + Softmax) Early-stage diabetes prediction Accuracy: High (> 90%) 5 An Effective Data Modeling Framework for Automatic Diabetes Prediction Patro et al. 2023 India PIMA Indian Diabetes dataset Deep CNN Data modeling for diabetes prediction Accuracy: 96.13% 6 Application of ensemble machine learning methods for diabetes diagnosis Ziyadullaev et al. Uzbekistan 2024 Pima Indian Diabetes dataset. Ensemble methods (Stacking) Analyze the use of ensemble ML for diabetes diagnosis Stacking model had higher accuracy compared to other methods 7 Classification of Diabetes Using Deep Learning and SVM Techniques K. Thaiyalnayaki Indian 2021 Pima Diabetes dataset Deep Learning Perceptron (DLP) and SVM Provide automatic classification of diabetes DLP classifier performed well with high accuracy 8 Deep Learning for Diabetes: A Systematic Review Zhu et al. United Kingdom 2021 Pima Indian Diabetes dataset (PID). Systematic review of DL methods Examine applications of DL in diabetes prediction DL methods significantly outperform traditional ML 9 Deep Learning-Based Glucose Prediction Models: A Guide for Practitioners and a Curated Dataset for Improved Diabetes Management Langarica et al. Chile 2024 The study collected ambulatory data from 20 participants, including healthy individuals and those with Type 1 Diabetes Mellitus (T1DM), to predict blood glucose levels. Multiple Recurrent Neural Networks (RNNs) Predict glucose levels (short- and long-term) DL methods produced lower error (better predictions) 10 Deep Unsupervised Machine Learning for Early Diabetes Risk Prediction using Ensemble Feature Selection and Deep Belief Neural Networks Olabanjo et al. Nigeria 2023 The study uses a diabetes dataset from Sylhet Diabetes Hospital in Bangladesh, containing data Deep Belief Network (DBN) Propose DBN for diabetes diagnosis DBN outperformed classical models in predictive accuracy 11 Diabetes detection based on machine learning and deep learning approaches Wee Malaysia 2023 Images data Convolutional Neural Networks (CNNs) trained on imaging data Develop predictive models for diabetes complications using imaging data High predictive accuracy; model driven by liver and pancreas features 12 Using deep learning to predict abdominal age from liver and pancreas magnetic resonance images. Alan et al. United States 2022 MRI images of liver and pancreas; genetic data; clinical and sociodemographic variables Deep learning-based segmentation (nnU-Net) Investigate fatty pancreas and imaging biomarkers for diabetes IPFD and fatty pancreas (FP) were significant predictors 13 Associations of Intrapancreatic Fat Deposition With Incident Diseases of the Exocrine and Endocrine Pancreas: A UK Biobank Prospective Cohort Study. Xiaowu et al. United Kingdom 2024 MRI imaging (quantified IPFD); clinical and demographic data Domain Knowledge-Infused CNN (DK-CNN) Develop a domain-knowledge-infused model for diabetes prediction DK-CNN retrieved the most clinically similar patients 14 Using similar patients to predict complication in patients with diabetes, hypertension, and lipid disorder: a domain knowledge-infused convolutional neural network approach. Oei et al. United Kingdom 2023 Electronic Health Records (EHRs) of 169,434 patients with diabetes Neural network models trained on retinal images Predict incident diabetes from retinal imaging Demonstrated feasibility of using retinal imaging for diabetes prediction 15 Prediction of hypertension, hyperglycemia and dyslipidemia from retinal fundus photographs via deep learning: A cross-sectional study of chronic diseases in central China. Zhang et al. United states 2020 Retinal fundus images (1,222 high-quality images from 625 subjects) Deep CNN with preprocessing Build an automatic system to classify diabetes type using ECG signals Accuracy: 99.56%, Precision: 99.72% 16 Association Between Aortic Imaging Features and Impaired Glucose Metabolism: A Deep Learning Population Phenotyping Approach. Rau et al. United states 2025 Magnetic Resonance Imaging dataset Deep learning models applied to retinal fundus photographs Develop AI models for detecting prediabetes and diabetes using retinal imaging AUC: 0.822 for prediabetes; 0.895 for diabetes detection 17 Association between depressive symptoms and diagnosis of diabetes and its complications: A network analysis in electronic health records. Wan et al. Switzerland 2022 "EHR notes of 52,139 inpatients diagnosed with T2DM. CNN models (VGG16, InceptionV3) Use transfer learning for diabetic retinopathy classification Accuracy up to 90.9% 18 Research progress on ocular complications caused by type 2 diabetes mellitus and the function of tears and blepharons. Wang et al. Poland 2024 Long Short-Term Memory (LSTM) Network Predict type 2 diabetes from longitudinal clinical data AUC: 0.83 19 Deep learning architectures for multi-label classification of intelligent health risk prediction. Maxwell et al. United Kingdom 2017 Deep CNN and hybrid model Develop hybrid model for early-stage diabetes prediction Accuracy: 96.3% 20 Deep learning to estimate impaired glucose metabolism from Magnetic Resonance Imaging of the liver: An opportunistic population screening approach. Michel et al. United States 2024 Magnetic Resonance Imaging (MRI) data from a prospective population study. CNN and Transfer Learning (DenseNet121) Classify diabetic retinopathy severity levels from retinal images Accuracy: 96.5% 21 Fully Automated Myocardial Strain Estimation from Cardiovascular MRI-tagged Images Using a Deep Learning Framework in the UK Biobank. Ferdian et al. Unted States 2020 Physical examination records of 110,300 anonymous patients. Deep Convolutional Neural Network (DCNN) Predict diabetes using fundus images AUC: 0.820 22 Detection Rate of Diabetic Retinopathy Before and After Implementation of Autonomous AI-based Fundus Photograph Analysis in a Resource-Limited Area in Belize. Esmaeilkhanian et al. New Zealand 2025 Magnetic Resonance Imaging (MRI) data from a prospective population study. Ensemble of CNN models (ResNet, DenseNet, Inception) Improve diabetes diagnosis via ensemble learning Accuracy: 94.5% 23 Diabetes Prediction Using Enhanced SVM and Deep Neural Network Learning Techniques: An Algorithmic Approach for Early Screening of Diabetes: Nagaraj et al. Indian 2021 Cardiac MRI-tagged images from the U.K. Biobank. CNN-based multi-task learning model Predict diabetic retinopathy and cardiovascular risk simultaneously High predictive performance reported 24 Diabetes Prediction Using Machine Learning and Flask Raju et al. Indian 2024 Patient records from the Stanford Belize Vision Clinic (SBVC). LSTM Network Early prediction of type 2 diabetes using longitudinal EMR data Accuracy: 91.2% 25 Diabetes Prediction: A Deep Learning Approach Ayon et al. Bangladesh 2019 Tabular patient data: 768 patient records with 8 features and a binary outcome (“Positive” or “Negative”) Transfer learning with VGG19 and ResNet50 Classify stages of diabetic retinopathy Accuracy: 85.6% (VGG19), 88.2% (ResNet50) 26 Deep Learning Model for Predicting Diabetes Disease Using SVM Anusuya et al. Singapore 2023 Tabular patient health data (likely with common features like age, BMI, glucose level, etc.) Deep learning CNN model Classify diabetic retinopathy severity using fundus images Accuracy: 94.3% 27 Machine Learning and Deep Learning Models for Nocturnal High- and Low-Glucose Prediction in Adults with Type 1 Diabetes Kozinetz et al. Russia 2024 Pima Indian Diabetes (PID) dataset CNN and Transfer Learning (ResNet101) Detect diabetic retinopathy from retinal images Accuracy: 92.6% 28 A Comparative analysis of Early Stage Diabetes Prediction using Machine Learning and Deep learning Approach Refat et al. Bangladesh. 2021 Pima Indian Diabetes Dataset CNN-LSTM hybrid model Predict blood glucose levels from sequential data RMSE values improved compared to baseline methods 29 A Deep Learning-based Architecture for Diabetes Detection, Prediction, and Classification Fakhar et al. Pakistan 2024 CGM data from 380 subjects with Type 1 Diabetes on Multiple Daily Injections (MDIs) 3D-CNN Model Predict type 2 diabetes using abdominal CT scans Dice coefficient: 0.69 30 Machine Learning and Deep Learning Techniques Applied to Diabetes Research: A Bibliometric Analysis García-Jaramillo et al. Colombia 2024 UCI Diabetes dataset with 17 attributes including the target class CNN with retinal image analysis Non-invasive prediction of diabetes using retina images AUC: 0.731 (algorithm alone), improved to 0.844 when combined with risk factors 31 Review Classification of Diabetes Using Machine Learning Technics Elias et al. Iraq 2024 Pima Indian Diabetes Dataset (PIDD) Application of machine learning algorithms to diabetes diagnosis Explore how ML algorithms classify diabetes cases and evaluate their performances ML methods assist in diagnosing diabetes; performance varies by algorithm 32 A Deep Learning-Based Diabetes Diagnosis Model on PIMA Image Dataset Avinash Bhagat, Tawseef Ahmed Teli Indian 2024 1773 scientific articles retrieved from Scopus database Preprocessing data to clean and transform into images, then CNN training Develop a novel deep learning approach for early diabetes diagnosis Achieved 97.19% accuracy 33 Performance Analysis of Deep Neural Network and Machine Learning Algorithms for Diabetes Prediction Tripathy et al. Indian 2023 Patient health data Applied multiple machine learning algorithms including deep neural networks Develop and compare various ML techniques for diabetes prediction Deep neural network outperformed traditional ML algorithms 34 Diabetes & Heart Disease Prediction Using Machine Learning Dhande et al. Indian 2022 PIMA Indians Diabetes Database Feature selection combined with ensemble learning (Voting Classifier) Propose an ensemble ML approach for diabetes and heart disease prediction Ensemble model achieved higher accuracy than individual models 35 Diabetes classification using MapReduce-based capsule network Arun et al. Indian 2024 PIMA Indians Diabetes Dataset Deep learning using Capsule Networks integrated with big data (MapReduce) Apply big data techniques with deep learning for diabetes classification CapsNet-MapReduce model outperformed conventional classifiers 36 Diabetes detection using deep learning algorithms G et al. Indian 2018 Clinical data for heart disease and diabetes Deep learning models (CNN, LSTM, CNN-LSTM hybrid) Develop a deep learning-based system for early diabetes detection Achieved 95.7% classification accuracy with CNN-LSTM model 37 REMOTE PATIENT MONITORING AND CLASSIFICATION OF DIABETES SUBTYPES CLASSIFICATION USING DEEP-LEARNING RECONSTRUCTION ALGORITHM Madurai et al. Indian 2024 Big data from patient health records Deep learning reconstruction using CNN and RNN Classify diabetes types remotely using IoT and DL Overall Accuracy: 92%; Type 1 Diabetes Sensitivity: 94% 38 Deep Learning of the Retina Enables Phenome- and Genome-Wide Analyses of the Microvasculature. Zekavat et al. United States 2022 Heart Rate Variability (HRV) signals derived from ECG Convolutional Neural Networks (CNNs) Explore retinal images for genetic and phenotypic analysis, including diabetes prediction Low vascular density/fractal dimension associated with diabetes 39 Nailfold capillaroscopy and deep learning in diabetes. Shah et al. Australia 2023 Continuous physiological data from remote monitoring devices CNNs applied to nailfold capillary images Detect diabetes based on vascular changes AUROC for diabetes detection: 0.84 (95% CI: 0.76–0.91) 40 Machine learning for predicting diabetes risk in western China adults. Li et al. China (Xinjiang region) 2023 Retinal fundus images (97,895 images from 54,813 participants) + phenotypic and genotypic data Integrated Learning using ensemble models like XGBoost Develop an ML model to predict diabetes risk XGBoost outperformed five other algorithms in accuracy 41 Artificial intelligence with temporal features outperforms machine learning in predicting diabetes. Naveed et al. Canada 2023 Nailfold capillary images from 120 participants (5236 images total) Deep learning models (including LSTM) compared with traditional ML Investigate and compare predictive capabilities for diabetes using temporal data Deep learning models outperformed traditional ML models in AUC 42 Stratification of diabetes in the context of comorbidities, using representation learning and topological data analysis. Wamil et al. United Kingdom 2023 Over 4 million national physical exam records: questionnaire, routine physical examination, and laboratory test indices Transformer-based model (BEHRT) to learn embeddings Validate a new model to stratify diabetes patients by cardiovascular risk Identified four diabetes phenotypes with distinct cardiovascular outcomes 43 Predicting the diabetic foot in the population of type 2 diabetes mellitus from tongue images and clinical information using multi-modal deep learning. Tian et al. China 2024 Electronic Medical Records (EMR) of 19,000 + patients ResNet-50 deep neural network for feature extraction Develop and validate a non-invasive diabetic foot prediction model Tongue features improved prediction; performance better with deep learning 44 Predictive models for posttransplant diabetes mellitus in kidney transplant recipients using machine learning and deep learning approach: a nationwide cohort study from South Korea. Choi et al. South Korea 2025 Clinical Practice Research Datalink (CPRD); 9,967 patients with new-onset diabetes XGBoost, CatBoost, LightGBM, Logistic Regression Predict risk of posttransplant diabetes mellitus (PTDM) XGBoost model performed best in AUC and precision metrics 45 Data-Driven Two-Stage Framework for Identification and Characterization of Different Antibiotic-Resistant Escherichia coli Isolates Based on Mass Spectrometry Data. Chung et al. Taiwan 2023 Clinical features, tongue images (traditional Chinese medicine), and biometric parameters of 391 patients Two-stage ML framework: peak-based feature extraction + XGBoost Develop a data-driven two-stage model for diabetes detection XGBoost achieved highest AUC for classification 46 Two-Stage Model-Based Predicting PV Generation with the Conjugation of IoT Sensor Data. Heo et al. Republic of Korea 2023 Pre- and post-transplant clinical variables (n = 72 features) from 3,213 kidney transplant recipients, using national transplant registry data Two-stage ML framework: ( 1 ) informative peak-based grouping using Random Forest importance, ( 2 ) prediction using Logistic Regression, SVM, RF, XGBoost ( 1 ) Predict future IoT sensor data using predicted environmental data; ( 2 ) Predict PV voltage using predicted IoT + environmental data using neural networks. Improved prediction accuracy by > 12% over baseline models. 47 Estimation of in-situ biogas upgrading in microbial electrolysis cells via direct electron transfer: Two-stage machine learning modeling based on a NARX-BP hybrid neural network. Xiao et al. China 2021 MALDI-TOF MS spectra of 37,918 E. coli isolates; AMR labels for five antibiotics (AMC, CAZ, CIP, CRO, CXM) Two-stage model: ( 1 ) Predict future IoT sensor data using predicted environmental data; ( 2 ) Predict PV voltage using predicted IoT + environmental data using neural networks Two-stage neural network model using NARX + Backpropagation hybrid networks to predict methane production in bioelectrochemical systems. R² = 0.918, Mean Error Square (MES) = 6.52 × 10⁻². 48 Predicting in-hospital length of stay: a two-stage modeling approach to account for highly skewed data. Xu et al. United States 2022 IoT sensor data, meteorological agency data (environmental data), PV voltage data Two-stage neural network model using Nonlinear AutoRegressive Two-stage hybrid classification-regression model for predicting in-hospital length of stay by classifying short vs. long stays, then regression for short stays. Good predictive accuracy for short stays; challenges for long stays. Specific accuracy values were not clearly stated. 49 MRI-based two-stage deep learning model for automatic detection and segmentation of brain metastases. Li et al. China 2023 MEC process data, including intermediate variables from biocathode microbial electrolysis for in-situ CH₄ upgrading Comparison of loss functions (mean squared error, mean absolute error, mean relative error), algorithms (LASSO, Random Forests, multilayer perceptron). Two-stage deep learning model for automatic detection and segmentation of brain metastases (segmentation + multi-scale classification). Sensitivity: 90%, Precision: 56%, Dice score: 81%, Relative Volume Difference (RVD): 20%. 50 Two-stage deep learning framework for occlusal crown depth image generation. Roh et al. Republic of Korea 2024 Electronic health record data on length of stay from elective surgeries. Development of a two-stage deep learning model consisting of a lightweight segmentation network for generating metastases proposals and a multi-scale classification network for false-positive suppression. Two-stage deep learning model for occlusal crown image generation (segmentation + GAN-based inpainting). Reduced MSE from 0.007001 to 0.002618; increased Dice score from 0.9333 to 0.964 The most commonly used dataset was the PIMA Indian Diabetes Dataset, which appeared in 10 (20%) of all studies, followed by the UCI Machine Learning Repository and other sources. Convolutional Neural Networks (CNNs) were the most frequently used model type, applied in 16 (38%) of the studies, followed by Long Short-Term Memory (LSTM) networks, used in approximately 10 (20%). The distribution of single-stage deep learning methods identified in the reviewed studies is summarized in Table 2 . Table 2 Distribution of single-stage deep learning methods used for diabetes classification and prediction Method Frequency Percentage (%) CNN 6 13.95 CNN + LSTM 5 11.62 CNN multitask 4 9.30 Capsule Network 3 6.98 CNN on nailfold image 3 6.98 Ensemble CNN 2 4.65 Deep CNN + Hybrid 2 4.65 ResNet (various) 2 4.65 CNN + RNN 2 4.65 CNN + DenseNet121 2 4.65 Others (frequency < 2 ) 12 27.91 Among the reviewed studies, only 7 (14%) implemented two-stage modeling frameworks, while the remaining 43 (86%) utilized single-stage models that performed classification or prediction within a unified analytical pipeline. 15 (30%) of the studies were conducted in 2024, and in the two-stage frameworks, models typically carried out an initial classification step before predicting subsequent outcomes, often focusing on diabetes-related complications such as retinopathy. The two-stage modelling approaches identified in the literature are summarized in Table 3 . Table 3 Distribution of two-stage deep learning approaches used in diabetes classification and prediction Method Frequency Percentage (%) DL + GAN 2 28.57 DL segmentation 2 28.57 Hybrid DL + ML 1 14.29 RF + XGBoost 1 14.29 DL + XGBoost 1 14.29 The distribution of deep learning models used for diabetes classification and prediction across the included studies is illustrated in Fig. 2 . The geographical distribution of studies included in this review is illustrated in Fig. 3 . Discussion This scoping review systematically mapped the application of deep learning and machine learning models for the classification and prediction of diabetes, synthesizing evidence from 50 studies published between 2017 and 2025. The review highlights the substantial growth and promise of deep learning techniques, particularly Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), in improving diagnostic accuracy for diabetes. CNNs emerged as the most frequently applied architecture, reflecting their strength in handling structured clinical and imaging data. Meanwhile, LSTM models demonstrated superior capabilities when handling longitudinal electronic medical record (EMR) data. Despite impressive advances in model accuracy, several limitations remained. Most studies relied heavily on publicly available datasets, such as the PIMA Indian Diabetes Dataset, with relatively few using real-world clinical datasets. Furthermore, the majority of models focused on binary classification (diabetic vs. non-diabetic), without extending to the nuanced classification of diabetes subtypes or the prediction of future complications. Only a small proportion (approximately 14%) of studies explored two-stage deep learning models, emphasizing a major gap in the literature. Two-stage models, which sequentially classify disease types and predict outcomes, are critical for personalized medicine but remain underexplored in diabetes research. Moreover, studies originating from Africa, including Nigeria and other Sub-Saharan countries, were extremely limited, raising concerns about the global generalizability of existing models. Limitations This review has some limitations. First, the study relied primarily on published literature retrieved from PubMed and Google Scholar, which may have excluded relevant studies indexed in other databases. Second, some studies included in the review demonstrated two-stage modeling frameworks in broader biomedical or predictive analytics contexts rather than being exclusively focused on diabetes datasets. These studies were included to illustrate methodological approaches relevant to sequential modeling and to highlight potential frameworks that could be adapted for diabetes classification and prediction in future research. Conclusion This scoping review systematically explored the current landscape of deep learning models for the classification and prediction of diabetes. The review identified a predominance of CNN- and RNN-based models, often achieving high levels of predictive accuracy. However, critical gaps remain, including the limited application of two-stage models, insufficient integration of multimodal datasets, a lack of African population-based studies, and minimal focus on model interpretability. These gaps underscore the urgent need for more robust, generalizable, and clinically applicable solutions in diabetes prediction and management. Declarations Ethics approval and consent to participate Not applicable. This study is a scoping review based exclusively on previously published literature and did not involve human participants or personal data. Consent for publication Not applicable. Funding The authors received no specific funding for this work. Competing interests The authors declare that they have no competing interests. References Bülbül MA (2024) A novel hybrid deep learning model for early stage diabetes risk prediction. J Supercomputing 80(13):19462–19484. https://doi.org/10.1007/s11227-024-06211-9 Chung C-R, Wang H-Y, Yao C-H, Wu L-C, Lu J-J, Horng J-T, Lee T-Y (2023) Data-Driven Two-Stage Framework for Identification and Characterization of Different Antibiotic-Resistant Escherichia coli Isolates Based on Mass Spectrometry Data. Microbiol Spectr 11(3):e0347922. https://doi.org/10.1128/spectrum.03479-22 El-Bashbishy AE-S, El-Bakry HM (2024) Pediatric diabetes prediction using deep learning. Sci Rep 14(1):4206. https://doi.org/10.1038/s41598-024-51438-4 G., S., R., V., & K.P., S. (2018) Diabetes detection using deep learning algorithms. ICT Express, 4 (4), 243–246. https://doi.org/10.1016/j.icte.2018.10.005 International Diabetes Federation (2025) IDF Diabetes Atlas (11th ed.). https://www.diabetesatlas.org Langarica S, De La Vega D, Cariman N, Miranda M, Andrade DC, Núñez F, Rodriguez-Fernandez M (2024) Deep Learning-Based Glucose Prediction Models: A Guide for Practitioners and a Curated Dataset for Improved Diabetes Management. IEEE Open J Eng Med Biology 5:467–475. https://doi.org/10.1109/OJEMB.2024.3365290 Olabanjo O, Wusu A, Mazzara M (2023) Deep Unsupervised Machine Learning for Early Diabetes Risk Prediction using Ensemble Feature Selection and Deep Belief Neural Networks. Phys Sci. https://doi.org/10.20944/preprints202301.0208.v1 Ong KL, Stafford LK, McLaughlin SA, Boyko EJ, Vollset SE, Smith AE, Dalton BE, Duprey J, Cruz JA, Hagins H, Lindstedt PA, Aali A, Abate YH, Abate MD, Abbasian M, Abbasi-Kangevari Z, Abbasi-Kangevari M, Abd ElHafeez S, Abd-Rabu R, Vos T (2023) Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050: A systematic analysis for the Global Burden of Disease Study 2021. Lancet 402(10397):203–234. https://doi.org/10.1016/S0140-6736(23)01301-6 Patro KK, Allam JP, Sanapala U, Marpu CK, Samee NA, Alabdulhafith M, Plawiak P (2023) An effective correlation-based data modeling framework for automatic diabetes prediction using machine and deep learning techniques. BMC Bioinformatics 24(1):372. https://doi.org/10.1186/s12859-023-05488-6 Tallam H, Elton DC, Lee S, Wakim P, Pickhardt PJ, Summers RM (2022) Fully Automated Abdominal CT Biomarkers for Type 2 Diabetes Using Deep Learning. Radiology 304(1):85–95. https://doi.org/10.1148/radiol.211914 Thaiyalnayaki K (2021) Classification of Diabetes Using Deep Learning and SVM Techniques. Int J Curr Res Rev 13(01):146–149. https://doi.org/10.31782/IJCRR.2021.13127 Tripathy N, Moharana B, Balabantaray SK, Nayak SK, Pati A, Panigrahi A (2024) A Comparative Analysis of Diabetes Prediction Using Machine Learning and Deep Learning Algorithms in Healthcare. 2024 International Conference on Advancements in Smart, Secure and Intelligent Computing (ASSIC) , 1–6. https://doi.org/10.1109/ASSIC60049.2024.10508008 Tripathy N, Das GC, Hota S, Navak SK, Satapathy P, Prusty S (2023) Performance Analysis of Deep Neural Network and Machine Learning Algorithms for Diabetes Prediction. 2023 2nd International Conference on Ambient Intelligence in Health Care (ICAIHC) , 1–6. https://doi.org/10.1109/ICAIHC59020.2023.10431434 Xu H, Zheng Q, Zhu J, Xie Z, Cheng H, Li P, Ji Y (2022) A Deep Learning Model Incorporating Knowledge Representation Vectors and Its Application in Diabetes Prediction. Disease Markers , 2022 , 1–17. https://doi.org/10.1155/2022/7593750 Yun J-S, Kim J, Jung S-H, Cha S-A, Ko S-H, Ahn Y-B, Won H-H, Sohn K-A, Kim D (2021) A Deep Learning Model for Screening Type 2 Diabetes from Retinal Photographs. Endocrinol (including Diabetes Mellitus Metabolic Disease). https://doi.org/10.1101/2021.06.29.21259606 Zhu T, Li K, Herrero P, Georgiou P (2021) Deep Learning for Diabetes: A Systematic Review. IEEE J Biomedical Health Inf 25(7):2744–2757. https://doi.org/10.1109/JBHI.2020.3040225 Ziyadullaev D, Muhamediyeva D, Madazimov K, Madazimov M, Temirov P, Abdukadirov D (2024) Application of ensemble machine learning methods for diabetes diagnosis. BIO Web of Conferences , 121 , 01002. https://doi.org/10.1051/bioconf/202412101002 World Health Organization (2024), April 12 Diabetes . https://www.who.int/news-room/fact-sheets/detail/diabetes Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9081794","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Systematic Review","associatedPublications":[],"authors":[{"id":603682711,"identity":"d2eae0d1-5a07-43d1-b461-59b930b96395","order_by":0,"name":"Adedayo Emmanuel OJO","email":"data:image/png;base64,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","orcid":"https://orcid.org/0009-0004-8938-8247","institution":"University of Ibadan, Ibadan, Nigeria","correspondingAuthor":true,"prefix":"","firstName":"Adedayo","middleName":"Emmanuel","lastName":"OJO","suffix":""},{"id":603682712,"identity":"accdadf5-ff03-4651-9344-2b963eb15791","order_by":1,"name":"Rotimi Felix AFOLABI","email":"","orcid":"","institution":"University of Ibadan, Ibadan, Nigeria","correspondingAuthor":false,"prefix":"","firstName":"Rotimi","middleName":"Felix","lastName":"AFOLABI","suffix":""},{"id":603682713,"identity":"7a9755df-a6b7-4fd3-b478-89074ef50379","order_by":2,"name":"William BALOGUN","email":"","orcid":"","institution":"University of Ibadan, Ibadan, Nigeria.","correspondingAuthor":false,"prefix":"","firstName":"William","middleName":"","lastName":"BALOGUN","suffix":""}],"badges":[],"createdAt":"2026-03-10 09:12:28","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":true,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":true},"doi":"10.21203/rs.3.rs-9081794/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9081794/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104392856,"identity":"32c2e8c5-1d65-4f48-9c79-bfdda65dd895","added_by":"auto","created_at":"2026-03-11 10:37:33","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":85594,"visible":true,"origin":"","legend":"\u003cp\u003ePRISMA-ScR flow diagram describing the study selection process.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9081794/v1/b3bdffb57c82373f12c77651.png"},{"id":104405891,"identity":"f21d9919-7639-461a-b8a4-38799997d83f","added_by":"auto","created_at":"2026-03-11 12:24:05","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":115684,"visible":true,"origin":"","legend":"\u003cp\u003eBubble plot of deep learning methods by country and year.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9081794/v1/ef13295bee9c98488c9ca0ce.png"},{"id":104392859,"identity":"75d08d3e-19c7-4d55-806d-b2e056bba5da","added_by":"auto","created_at":"2026-03-11 10:37:33","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":309335,"visible":true,"origin":"","legend":"\u003cp\u003eGlobal distribution of deep learning models used in diabetes research.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9081794/v1/947363ac54754ab6792ae5f4.png"},{"id":104409533,"identity":"6f25ca38-fabd-4ebe-9f32-69d277b60db4","added_by":"auto","created_at":"2026-03-11 12:45:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1424631,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9081794/v1/0dcb87bc-70f0-4782-aaa4-d0267e93f89c.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eScoping Review on Deep Learning Model for Classification and Prediction of Diabetes\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDiabetes mellitus is a chronic metabolic disorder characterized by persistent hyperglycemia, resulting from defects in insulin secretion, insulin action, or both (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). In 2021, it accounted for more than 2\u0026nbsp;million deaths globally and is a major cause of complications such as blindness, kidney failure, heart attacks, stroke, and diabetic foot ulcer (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). As of 2024, global prevalence is 11.1% (588.7\u0026nbsp;million people), and about 45.8% remain undiagnosed (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). In Nigeria, prevalence among adults is estimated at 3.7% (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Type 2 diabetes accounts for more than 90% of all cases (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). Traditional diagnostics such as HbA1c and fasting glucose often fail to detect (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eArtificial intelligence (AI) has gained increasing attention in medical diagnostics due to its ability to analyze complex and high-dimensional clinical data beyond the capacity of traditional diagnostic approaches (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Conventional methods for diagnosing chronic noncommunicable diseases, such as diabetes mellitus, often depend on fixed clinical thresholds and expert interpretation, which may limit early detection and consistency across populations (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). AI-based techniques, particularly machine learning and deep learning models, offer a complementary approach by learning hidden patterns from biomedical data and integrating multiple risk factors simultaneously. Moreover, recent advances seek to improve clinical relevance by incorporating medical knowledge into AI models, addressing concerns related to interpretability and trust. As a result, AI presents a promising framework for supporting early diagnosis, risk assessment, and decision-making in the management of diabetes mellitus. (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Deep learning leverages multilayer neural networks to discover patterns in structured health data (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePrevious studies have suggested the use of one-stage DL models, performing prediction or classification in a single step. Two-stage models classify diabetes and then predict outcomes; they are easier to interpret (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). The need for sequential, clinically interpretable models has led to growing interest in Two-stage DL architectures, where the first stage performs classification (Type 1 vs. Type 2) and the second stage predicts complications or disease progression.\u003c/p\u003e \u003cp\u003eThis scoping review aimed to explore and map the existing literature on DL models for diabetes classification and prediction. In doing so, it seeks to identify research gaps and provide direction for future studies, particularly those addressing local contexts in underserved regions.\u003c/p\u003e"},{"header":"Methodology","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design\u003c/h2\u003e \u003cp\u003e A scoping review was conducted using the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines to map existing evidence on the use of deep learning models for diabetes classification and prediction.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eResearch Question\u003c/h3\u003e\n\u003cp\u003eWhat deep learning approaches have been used to classify Type 1 and Type 2 diabetes, and how are they applied and evaluated in existing literature\u003c/p\u003e\n\u003ch3\u003eSearch Strategy\u003c/h3\u003e\n\u003cp\u003eA systematic search was performed in PubMed and Google Scholar databases using Boolean search operators: \"deep learning\" AND \"model\" AND \"diabetes\". The search included studies applying deep learning to the classification and/or prediction of diabetes, without restriction on publication date, language, country, or population ethnicity. The process of study identification, screening, eligibility assessment, and final inclusion of studies is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\n\u003ch3\u003eEligibility Criteria\u003c/h3\u003e\n\u003cp\u003eStudies were eligible for inclusion if they involved individuals diagnosed with Type 1 or Type 2 diabetes mellitus, applied deep learning or hybrid artificial intelligence models for the classification and/or prediction of diabetes, and were conducted within healthcare or diagnostic contexts globally. Only studies published between January 2017 and August 2025 were considered. Studies were excluded if they focused exclusively on traditional machine learning methods without deep learning components, involved non-human subjects, or addressed disease contexts unrelated to diabetes mellitus. In addition, some studies demonstrating two-stage modeling frameworks in biomedical or predictive analytics contexts were reviewed to illustrate methodological approaches relevant to diabetes prediction.\u003c/p\u003e\n\u003ch3\u003eStudy Selection and Screening\u003c/h3\u003e\n\u003cp\u003eInitial records were screened for duplicates, followed by title and abstract screening. Eligible full-text articles were assessed for inclusion. Screening and extraction were performed by two independent reviewers.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eData Extraction\u003c/h2\u003e \u003cp\u003eData extraction was conducted using a structured Excel spreadsheet to systematically chart key study characteristics. Extracted information included the author(s), year of publication, country of study, the dataset used, the type of deep learning model applied, whether a one-stage or two-stage classification approach was employed, and the reported performance metrics.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eA total of 750 articles were identified. After removing 97 duplicates and screening titles and abstracts, 248 articles were retrieved for full-text review. Of these, 198 articles were excluded based on the eligibility criteria. A total of 50 studies met the inclusion criteria and were included in this scoping review. The studies varied in terms of data sources, deep learning architectures, and evaluation metrics used for diabetes classification and prediction. The characteristics of the included studies are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCharacteristics of studies included in the scoping review.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eS/N\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTitle\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAuthors\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eYear\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDataset used\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eModel/Method\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eKey Focus\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003ePerformance Metrics\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFully Automated Abdominal CT Biomarkers for Type Diabetes Using Deep Learning\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTallam et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eUnited States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCT imaging and clinical data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3D U-Net\u0026thinsp;+\u0026thinsp;Logistic Regression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eCT biomarkers for diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDice Score 0.69, AUC (Not specified)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA Deep Learning Model for Screening Type 2 Diabetes from Retinal Photographs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYun et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eUnited States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eRetinal images (UK Biobank)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eResNet18 fine-tuned\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eRetinal-based diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAUC 73% (algorithm), 84% (with traditional risk factors)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA Deep Learning Model Incorporating Knowledge Representation Vectors for Diabetes Prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eXu et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eClinical examination data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTHSAC model (self-attention\u0026thinsp;+\u0026thinsp;CNN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eKnowledge-based deep learning for diabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy (Not specified), F1-score high\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA novel Hybrid Deep Learning Model for Early-Stage Diabetes Risk Prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eB\u0026uuml;lb\u0026uuml;l\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTurkey\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eUCI Diabetes dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHybrid deep learning model (GA\u0026thinsp;+\u0026thinsp;Autoencoder\u0026thinsp;+\u0026thinsp;Softmax)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eEarly-stage diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: High (\u0026gt;\u0026thinsp;90%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAn Effective Data Modeling Framework for Automatic Diabetes Prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePatro et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eIndia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePIMA Indian Diabetes dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep CNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eData modeling for diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 96.13%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eApplication of ensemble machine learning methods for diabetes diagnosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eZiyadullaev et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUzbekistan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePima Indian Diabetes dataset.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEnsemble methods (Stacking)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eAnalyze the use of ensemble ML for diabetes diagnosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eStacking model had higher accuracy compared to other methods\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClassification of Diabetes Using Deep Learning and SVM Techniques\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eK. Thaiyalnayaki\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePima Diabetes dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep Learning Perceptron (DLP) and SVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eProvide automatic classification of diabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDLP classifier performed well with high accuracy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep Learning for Diabetes: A Systematic Review\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eZhu et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited Kingdom\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePima Indian Diabetes dataset (PID).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSystematic review of DL methods\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eExamine applications of DL in diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDL methods significantly outperform traditional ML\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep Learning-Based Glucose Prediction Models: A Guide for Practitioners and a Curated Dataset for Improved Diabetes Management\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLangarica et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eThe study collected ambulatory data from 20 participants, including healthy individuals and those with Type 1 Diabetes Mellitus (T1DM), to predict blood glucose levels.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMultiple Recurrent Neural Networks (RNNs)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict glucose levels (short- and long-term)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDL methods produced lower error (better predictions)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep Unsupervised Machine Learning for Early Diabetes Risk Prediction using Ensemble Feature Selection and Deep Belief Neural Networks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOlabanjo et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNigeria\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eThe study uses a diabetes dataset from Sylhet Diabetes Hospital in Bangladesh, containing data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep Belief Network (DBN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePropose DBN for diabetes diagnosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDBN outperformed classical models in predictive accuracy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiabetes detection based on machine learning and deep learning approaches\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWee\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMalaysia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eImages data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eConvolutional Neural Networks (CNNs) trained on imaging data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop predictive models for diabetes complications using imaging data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eHigh predictive accuracy; model driven by liver and pancreas features\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUsing deep learning to predict abdominal age from liver and pancreas magnetic resonance images.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAlan et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMRI images of liver and pancreas; genetic data; clinical and sociodemographic variables\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning-based segmentation (nnU-Net)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eInvestigate fatty pancreas and imaging biomarkers for diabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eIPFD and fatty pancreas (FP) were significant predictors\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssociations of Intrapancreatic Fat Deposition With Incident Diseases of the Exocrine and Endocrine Pancreas: A UK Biobank Prospective Cohort Study.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eXiaowu et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited Kingdom\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMRI imaging (quantified IPFD); clinical and demographic data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDomain Knowledge-Infused CNN (DK-CNN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop a domain-knowledge-infused model for diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDK-CNN retrieved the most clinically similar patients\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUsing similar patients to predict complication in patients with diabetes, hypertension, and lipid disorder: a domain knowledge-infused convolutional neural network approach.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOei et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited Kingdom\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eElectronic Health Records (EHRs) of 169,434 patients with diabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNeural network models trained on retinal images\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict incident diabetes from retinal imaging\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDemonstrated feasibility of using retinal imaging for diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePrediction of hypertension, hyperglycemia and dyslipidemia from retinal fundus photographs via deep learning: A cross-sectional study of chronic diseases in central China.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eZhang et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited states\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eRetinal fundus images (1,222 high-quality images from 625 subjects)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep CNN with preprocessing\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eBuild an automatic system to classify diabetes type using ECG signals\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 99.56%, Precision: 99.72%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssociation Between Aortic Imaging Features and Impaired Glucose Metabolism: A Deep Learning Population Phenotyping Approach.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRau et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited states\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMagnetic Resonance Imaging dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning models applied to retinal fundus photographs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop AI models for detecting prediabetes and diabetes using retinal imaging\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAUC: 0.822 for prediabetes; 0.895 for diabetes detection\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssociation between depressive symptoms and diagnosis of diabetes and its complications: A network analysis in electronic health records.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWan et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSwitzerland\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\"EHR notes of 52,139 inpatients diagnosed with T2DM.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCNN models (VGG16, InceptionV3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eUse transfer learning for diabetic retinopathy classification\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy up to 90.9%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eResearch progress on ocular complications caused by type 2 diabetes mellitus and the function of tears and blepharons.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWang et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePoland\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eLong Short-Term Memory (LSTM) Network\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict type 2 diabetes from longitudinal clinical data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAUC: 0.83\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep learning architectures for multi-label classification of intelligent health risk prediction.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMaxwell et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited Kingdom\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep CNN and hybrid model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop hybrid model for early-stage diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 96.3%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep learning to estimate impaired glucose metabolism from Magnetic Resonance Imaging of the liver: An opportunistic population screening approach.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMichel et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMagnetic Resonance Imaging (MRI) data from a prospective population study.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCNN and Transfer Learning (DenseNet121)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eClassify diabetic retinopathy severity levels from retinal images\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 96.5%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFully Automated Myocardial Strain Estimation from Cardiovascular MRI-tagged Images Using a Deep Learning Framework in the UK Biobank.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFerdian et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnted States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePhysical examination records of 110,300 anonymous patients.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep Convolutional Neural Network (DCNN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict diabetes using fundus images\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAUC: 0.820\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDetection Rate of Diabetic Retinopathy Before and After Implementation of Autonomous AI-based Fundus Photograph Analysis in a Resource-Limited Area in Belize.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEsmaeilkhanian et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNew Zealand\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMagnetic Resonance Imaging (MRI) data from a prospective population study.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEnsemble of CNN models (ResNet, DenseNet, Inception)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eImprove diabetes diagnosis via ensemble learning\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 94.5%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiabetes Prediction Using Enhanced SVM and Deep Neural Network Learning Techniques: An Algorithmic Approach for Early Screening of Diabetes:\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNagaraj et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCardiac MRI-tagged images from the U.K. Biobank.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCNN-based multi-task learning model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict diabetic retinopathy and cardiovascular risk simultaneously\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eHigh predictive performance reported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiabetes Prediction Using Machine Learning and Flask\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRaju et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePatient records from the Stanford Belize Vision Clinic (SBVC).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eLSTM Network\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eEarly prediction of type 2 diabetes using longitudinal EMR data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 91.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiabetes Prediction: A Deep Learning Approach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAyon et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBangladesh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTabular patient data: 768 patient records with 8 features and a binary outcome (\u0026ldquo;Positive\u0026rdquo; or \u0026ldquo;Negative\u0026rdquo;)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTransfer learning with VGG19 and ResNet50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eClassify stages of diabetic retinopathy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 85.6% (VGG19), 88.2% (ResNet50)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep Learning Model for Predicting Diabetes Disease Using SVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAnusuya et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSingapore\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTabular patient health data (likely with common features like age, BMI, glucose level, etc.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning CNN model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eClassify diabetic retinopathy severity using fundus images\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 94.3%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMachine Learning and Deep Learning Models for Nocturnal High- and Low-Glucose Prediction in Adults with Type 1 Diabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKozinetz et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRussia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePima Indian Diabetes (PID) dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCNN and Transfer Learning (ResNet101)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDetect diabetic retinopathy from retinal images\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAccuracy: 92.6%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA Comparative analysis of Early Stage Diabetes Prediction using Machine Learning and Deep learning Approach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRefat et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBangladesh.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePima Indian Diabetes Dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCNN-LSTM hybrid model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict blood glucose levels from sequential data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eRMSE values improved compared to baseline methods\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA Deep Learning-based Architecture for Diabetes Detection, Prediction, and Classification\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFakhar et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePakistan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCGM data from 380 subjects with Type 1 Diabetes on Multiple Daily Injections (MDIs)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3D-CNN Model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict type 2 diabetes using abdominal CT scans\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDice coefficient: 0.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMachine Learning and Deep Learning Techniques Applied to Diabetes Research: A Bibliometric Analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGarc\u0026iacute;a-Jaramillo et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eColombia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eUCI Diabetes dataset with 17 attributes including the target class\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCNN with retinal image analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eNon-invasive prediction of diabetes using retina images\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAUC: 0.731 (algorithm alone), improved to 0.844 when combined with risk factors\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReview Classification of Diabetes Using Machine Learning Technics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eElias et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIraq\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePima Indian Diabetes Dataset (PIDD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eApplication of machine learning algorithms to diabetes diagnosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eExplore how ML algorithms classify diabetes cases and evaluate their performances\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eML methods assist in diagnosing diabetes; performance varies by algorithm\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA Deep Learning-Based Diabetes Diagnosis Model on PIMA Image Dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAvinash Bhagat, Tawseef Ahmed Teli\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1773 scientific articles retrieved from Scopus database\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePreprocessing data to clean and transform into images, then CNN training\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop a novel deep learning approach for early diabetes diagnosis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAchieved 97.19% accuracy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePerformance Analysis of Deep Neural Network and Machine Learning Algorithms for Diabetes Prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTripathy et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePatient health data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eApplied multiple machine learning algorithms including deep neural networks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop and compare various ML techniques for diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDeep neural network outperformed traditional ML algorithms\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiabetes \u0026amp; Heart Disease Prediction Using Machine Learning\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDhande et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePIMA Indians Diabetes Database\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eFeature selection combined with ensemble learning (Voting Classifier)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePropose an ensemble ML approach for diabetes and heart disease prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eEnsemble model achieved higher accuracy than individual models\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiabetes classification using MapReduce-based capsule network\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eArun et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePIMA Indians Diabetes Dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning using Capsule Networks integrated with big data (MapReduce)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eApply big data techniques with deep learning for diabetes classification\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eCapsNet-MapReduce model outperformed conventional classifiers\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiabetes detection using deep learning algorithms\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eG et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eClinical data for heart disease and diabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning models (CNN, LSTM, CNN-LSTM hybrid)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop a deep learning-based system for early diabetes detection\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAchieved 95.7% classification accuracy with CNN-LSTM model\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eREMOTE PATIENT MONITORING AND CLASSIFICATION OF DIABETES SUBTYPES CLASSIFICATION USING DEEP-LEARNING RECONSTRUCTION ALGORITHM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMadurai et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eBig data from patient health records\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning reconstruction using CNN and RNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eClassify diabetes types remotely using IoT and DL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eOverall Accuracy: 92%; Type 1 Diabetes Sensitivity: 94%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep Learning of the Retina Enables Phenome- and Genome-Wide Analyses of the Microvasculature.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eZekavat et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHeart Rate Variability (HRV) signals derived from ECG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eConvolutional Neural Networks (CNNs)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eExplore retinal images for genetic and phenotypic analysis, including diabetes prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eLow vascular density/fractal dimension associated with diabetes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNailfold capillaroscopy and deep learning in diabetes.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eShah et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAustralia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eContinuous physiological data from remote monitoring devices\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCNNs applied to nailfold capillary images\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDetect diabetes based on vascular changes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAUROC for diabetes detection: 0.84 (95% CI: 0.76\u0026ndash;0.91)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMachine learning for predicting diabetes risk in western China adults.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLi et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina (Xinjiang region)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eRetinal fundus images (97,895 images from 54,813 participants) + phenotypic and genotypic data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eIntegrated Learning using ensemble models like XGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop an ML model to predict diabetes risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eXGBoost outperformed five other algorithms in accuracy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArtificial intelligence with temporal features outperforms machine learning in predicting diabetes.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNaveed et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCanada\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNailfold capillary images from 120 participants (5236 images total)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning models (including LSTM) compared with traditional ML\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eInvestigate and compare predictive capabilities for diabetes using temporal data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eDeep learning models outperformed traditional ML models in AUC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStratification of diabetes in the context of comorbidities, using representation learning and topological data analysis.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWamil et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited Kingdom\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eOver 4\u0026nbsp;million national physical exam records: questionnaire, routine physical examination, and laboratory test indices\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTransformer-based model (BEHRT) to learn embeddings\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eValidate a new model to stratify diabetes patients by cardiovascular risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eIdentified four diabetes phenotypes with distinct cardiovascular outcomes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePredicting the diabetic foot in the population of type 2 diabetes mellitus from tongue images and clinical information using multi-modal deep learning.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTian et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eElectronic Medical Records (EMR) of 19,000\u0026thinsp;+\u0026thinsp;patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eResNet-50 deep neural network for feature extraction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop and validate a non-invasive diabetic foot prediction model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTongue features improved prediction; performance better with deep learning\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePredictive models for posttransplant diabetes mellitus in kidney transplant recipients using machine learning and deep learning approach: a nationwide cohort study from South Korea.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eChoi et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSouth Korea\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eClinical Practice Research Datalink (CPRD); 9,967 patients with new-onset diabetes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eXGBoost, CatBoost, LightGBM, Logistic Regression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePredict risk of posttransplant diabetes mellitus (PTDM)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eXGBoost model performed best in AUC and precision metrics\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eData-Driven Two-Stage Framework for Identification and Characterization of Different Antibiotic-Resistant Escherichia coli Isolates Based on Mass Spectrometry Data.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eChung et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTaiwan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eClinical features, tongue images (traditional Chinese medicine), and biometric parameters of 391 patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTwo-stage ML framework: peak-based feature extraction\u0026thinsp;+\u0026thinsp;XGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eDevelop a data-driven two-stage model for diabetes detection\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eXGBoost achieved highest AUC for classification\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTwo-Stage Model-Based Predicting PV Generation with the Conjugation of IoT Sensor Data.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHeo et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRepublic of Korea\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePre- and post-transplant clinical variables (n\u0026thinsp;=\u0026thinsp;72 features) from 3,213 kidney transplant recipients, using national transplant registry data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTwo-stage ML framework: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) informative peak-based grouping using Random Forest importance, (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) prediction using Logistic Regression, SVM, RF, XGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) Predict future IoT sensor data using predicted environmental data; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) Predict PV voltage using predicted IoT\u0026thinsp;+\u0026thinsp;environmental data using neural networks.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eImproved prediction accuracy by \u0026gt;\u0026thinsp;12% over baseline models.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEstimation of in-situ biogas upgrading in microbial electrolysis cells via direct electron transfer: Two-stage machine learning modeling based on a NARX-BP hybrid neural network.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eXiao et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMALDI-TOF MS spectra of 37,918 E. coli isolates; AMR labels for five antibiotics (AMC, CAZ, CIP, CRO, CXM)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTwo-stage model: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) Predict future IoT sensor data using predicted environmental data; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) Predict PV voltage using predicted IoT\u0026thinsp;+\u0026thinsp;environmental data using neural networks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eTwo-stage neural network model using NARX\u0026thinsp;+\u0026thinsp;Backpropagation hybrid networks to predict methane production in bioelectrochemical systems.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eR\u0026sup2; = 0.918, Mean Error Square (MES)\u0026thinsp;=\u0026thinsp;6.52 \u0026times; 10⁻\u0026sup2;.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePredicting in-hospital length of stay: a two-stage modeling approach to account for highly skewed data.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eXu et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUnited States\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eIoT sensor data, meteorological agency data (environmental data), PV voltage data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTwo-stage neural network model using Nonlinear AutoRegressive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eTwo-stage hybrid classification-regression model for predicting in-hospital length of stay by classifying short vs. long stays, then regression for short stays.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eGood predictive accuracy for short stays; challenges for long stays. Specific accuracy values were not clearly stated.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMRI-based two-stage deep learning model for automatic detection and segmentation of brain metastases.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLi et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMEC process data, including intermediate variables from biocathode microbial electrolysis for in-situ CH₄ upgrading\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eComparison of loss functions (mean squared error, mean absolute error, mean relative error), algorithms (LASSO, Random Forests, multilayer perceptron).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eTwo-stage deep learning model for automatic detection and segmentation of brain metastases (segmentation\u0026thinsp;+\u0026thinsp;multi-scale classification).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eSensitivity: 90%, Precision: 56%, Dice score: 81%, Relative Volume Difference (RVD): 20%.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTwo-stage deep learning framework for occlusal crown depth image generation.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRoh et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRepublic of Korea\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eElectronic health record data on length of stay from elective surgeries.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDevelopment of a two-stage deep learning model consisting of a lightweight segmentation network for generating metastases proposals and a multi-scale classification network for false-positive suppression.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eTwo-stage deep learning model for occlusal crown image generation (segmentation\u0026thinsp;+\u0026thinsp;GAN-based inpainting).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eReduced MSE from 0.007001 to 0.002618; increased Dice score from 0.9333 to 0.964\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe most commonly used dataset was the PIMA Indian Diabetes Dataset, which appeared in 10 (20%) of all studies, followed by the UCI Machine Learning Repository and other sources. Convolutional Neural Networks (CNNs) were the most frequently used model type, applied in 16 (38%) of the studies, followed by Long Short-Term Memory (LSTM) networks, used in approximately 10 (20%). The distribution of single-stage deep learning methods identified in the reviewed studies is summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDistribution of single-stage deep learning methods used for diabetes classification and prediction\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMethod\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFrequency\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePercentage (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13.95\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCNN\u0026thinsp;+\u0026thinsp;LSTM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e11.62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCNN multitask\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9.30\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCapsule Network\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6.98\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCNN on nailfold image\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6.98\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEnsemble CNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.65\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep CNN\u0026thinsp;+\u0026thinsp;Hybrid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.65\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResNet (various)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.65\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCNN\u0026thinsp;+\u0026thinsp;RNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.65\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCNN\u0026thinsp;+\u0026thinsp;DenseNet121\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.65\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOthers (frequency\u0026thinsp;\u0026lt;\u0026thinsp;2 )\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e27.91\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAmong the reviewed studies, only 7 (14%) implemented two-stage modeling frameworks, while the remaining 43 (86%) utilized single-stage models that performed classification or prediction within a unified analytical pipeline. 15 (30%) of the studies were conducted in 2024, and in the two-stage frameworks, models typically carried out an initial classification step before predicting subsequent outcomes, often focusing on diabetes-related complications such as retinopathy. The two-stage modelling approaches identified in the literature are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDistribution of two-stage deep learning approaches used in diabetes classification and prediction\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMethod\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFrequency\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePercentage (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDL\u0026thinsp;+\u0026thinsp;GAN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e28.57\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDL segmentation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e28.57\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHybrid DL\u0026thinsp;+\u0026thinsp;ML\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e14.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRF\u0026thinsp;+\u0026thinsp;XGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e14.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDL\u0026thinsp;+\u0026thinsp;XGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e14.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe distribution of deep learning models used for diabetes classification and prediction across the included studies is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The geographical distribution of studies included in this review is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e This scoping review systematically mapped the application of deep learning and machine learning models for the classification and prediction of diabetes, synthesizing evidence from 50 studies published between 2017 and 2025.\u003c/p\u003e \u003cp\u003eThe review highlights the substantial growth and promise of deep learning techniques, particularly Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), in improving diagnostic accuracy for diabetes. CNNs emerged as the most frequently applied architecture, reflecting their strength in handling structured clinical and imaging data. Meanwhile, LSTM models demonstrated superior capabilities when handling longitudinal electronic medical record (EMR) data.\u003c/p\u003e \u003cp\u003eDespite impressive advances in model accuracy, several limitations remained. Most studies relied heavily on publicly available datasets, such as the PIMA Indian Diabetes Dataset, with relatively few using real-world clinical datasets. Furthermore, the majority of models focused on binary classification (diabetic vs. non-diabetic), without extending to the nuanced classification of diabetes subtypes or the prediction of future complications.\u003c/p\u003e \u003cp\u003eOnly a small proportion (approximately 14%) of studies explored two-stage deep learning models, emphasizing a major gap in the literature. Two-stage models, which sequentially classify disease types and predict outcomes, are critical for personalized medicine but remain underexplored in diabetes research. Moreover, studies originating from Africa, including Nigeria and other Sub-Saharan countries, were extremely limited, raising concerns about the global generalizability of existing models.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eThis review has some limitations. First, the study relied primarily on published literature retrieved from PubMed and Google Scholar, which may have excluded relevant studies indexed in other databases. Second, some studies included in the review demonstrated two-stage modeling frameworks in broader biomedical or predictive analytics contexts rather than being exclusively focused on diabetes datasets. These studies were included to illustrate methodological approaches relevant to sequential modeling and to highlight potential frameworks that could be adapted for diabetes classification and prediction in future research.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis scoping review systematically explored the current landscape of deep learning models for the classification and prediction of diabetes. The review identified a predominance of CNN- and RNN-based models, often achieving high levels of predictive accuracy.\u003c/p\u003e \u003cp\u003eHowever, critical gaps remain, including the limited application of two-stage models, insufficient integration of multimodal datasets, a lack of African population-based studies, and minimal focus on model interpretability. These gaps underscore the urgent need for more robust, generalizable, and clinically applicable solutions in diabetes prediction and management.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable. This study is a scoping review based exclusively on previously published literature and did not involve human participants or personal data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors received no specific funding for this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eB\u0026uuml;lb\u0026uuml;l MA (2024) A novel hybrid deep learning model for early stage diabetes risk prediction. J Supercomputing 80(13):19462\u0026ndash;19484. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11227-024-06211-9\u003c/span\u003e\u003cspan address=\"10.1007/s11227-024-06211-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChung C-R, Wang H-Y, Yao C-H, Wu L-C, Lu J-J, Horng J-T, Lee T-Y (2023) Data-Driven Two-Stage Framework for Identification and Characterization of Different Antibiotic-Resistant Escherichia coli Isolates Based on Mass Spectrometry Data. Microbiol Spectr 11(3):e0347922. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1128/spectrum.03479-22\u003c/span\u003e\u003cspan address=\"10.1128/spectrum.03479-22\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEl-Bashbishy AE-S, El-Bakry HM (2024) Pediatric diabetes prediction using deep learning. Sci Rep 14(1):4206. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-024-51438-4\u003c/span\u003e\u003cspan address=\"10.1038/s41598-024-51438-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG., S., R., V., \u0026amp; K.P., S. (2018) Diabetes detection using deep learning algorithms. ICT Express, \u003cem\u003e4\u003c/em\u003e(4), 243\u0026ndash;246. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.icte.2018.10.005\u003c/span\u003e\u003cspan address=\"10.1016/j.icte.2018.10.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInternational Diabetes Federation (2025) IDF Diabetes Atlas (11th ed.). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.diabetesatlas.org\u003c/span\u003e\u003cspan address=\"https://www.diabetesatlas.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLangarica S, De La Vega D, Cariman N, Miranda M, Andrade DC, N\u0026uacute;\u0026ntilde;ez F, Rodriguez-Fernandez M (2024) Deep Learning-Based Glucose Prediction Models: A Guide for Practitioners and a Curated Dataset for Improved Diabetes Management. IEEE Open J Eng Med Biology 5:467\u0026ndash;475. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/OJEMB.2024.3365290\u003c/span\u003e\u003cspan address=\"10.1109/OJEMB.2024.3365290\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOlabanjo O, Wusu A, Mazzara M (2023) Deep Unsupervised Machine Learning for Early Diabetes Risk Prediction using Ensemble Feature Selection and Deep Belief Neural Networks. Phys Sci. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.20944/preprints202301.0208.v1\u003c/span\u003e\u003cspan address=\"10.20944/preprints202301.0208.v1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOng KL, Stafford LK, McLaughlin SA, Boyko EJ, Vollset SE, Smith AE, Dalton BE, Duprey J, Cruz JA, Hagins H, Lindstedt PA, Aali A, Abate YH, Abate MD, Abbasian M, Abbasi-Kangevari Z, Abbasi-Kangevari M, Abd ElHafeez S, Abd-Rabu R, Vos T (2023) Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050: A systematic analysis for the Global Burden of Disease Study 2021. Lancet 402(10397):203\u0026ndash;234. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0140-6736(23)01301-6\u003c/span\u003e\u003cspan address=\"10.1016/S0140-6736(23)01301-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatro KK, Allam JP, Sanapala U, Marpu CK, Samee NA, Alabdulhafith M, Plawiak P (2023) An effective correlation-based data modeling framework for automatic diabetes prediction using machine and deep learning techniques. BMC Bioinformatics 24(1):372. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12859-023-05488-6\u003c/span\u003e\u003cspan address=\"10.1186/s12859-023-05488-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTallam H, Elton DC, Lee S, Wakim P, Pickhardt PJ, Summers RM (2022) Fully Automated Abdominal CT Biomarkers for Type 2 Diabetes Using Deep Learning. Radiology 304(1):85\u0026ndash;95. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1148/radiol.211914\u003c/span\u003e\u003cspan address=\"10.1148/radiol.211914\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThaiyalnayaki K (2021) Classification of Diabetes Using Deep Learning and SVM Techniques. Int J Curr Res Rev 13(01):146\u0026ndash;149. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.31782/IJCRR.2021.13127\u003c/span\u003e\u003cspan address=\"10.31782/IJCRR.2021.13127\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTripathy N, Moharana B, Balabantaray SK, Nayak SK, Pati A, Panigrahi A (2024) A Comparative Analysis of Diabetes Prediction Using Machine Learning and Deep Learning Algorithms in Healthcare. \u003cem\u003e2024 International Conference on Advancements in Smart, Secure and Intelligent Computing (ASSIC)\u003c/em\u003e, 1\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ASSIC60049.2024.10508008\u003c/span\u003e\u003cspan address=\"10.1109/ASSIC60049.2024.10508008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTripathy N, Das GC, Hota S, Navak SK, Satapathy P, Prusty S (2023) Performance Analysis of Deep Neural Network and Machine Learning Algorithms for Diabetes Prediction. \u003cem\u003e2023 2nd International Conference on Ambient Intelligence in Health Care (ICAIHC)\u003c/em\u003e, 1\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ICAIHC59020.2023.10431434\u003c/span\u003e\u003cspan address=\"10.1109/ICAIHC59020.2023.10431434\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu H, Zheng Q, Zhu J, Xie Z, Cheng H, Li P, Ji Y (2022) A Deep Learning Model Incorporating Knowledge Representation Vectors and Its Application in Diabetes Prediction. \u003cem\u003eDisease Markers\u003c/em\u003e, \u003cem\u003e2022\u003c/em\u003e, 1\u0026ndash;17. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1155/2022/7593750\u003c/span\u003e\u003cspan address=\"10.1155/2022/7593750\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYun J-S, Kim J, Jung S-H, Cha S-A, Ko S-H, Ahn Y-B, Won H-H, Sohn K-A, Kim D (2021) A Deep Learning Model for Screening Type 2 Diabetes from Retinal Photographs. Endocrinol (including Diabetes Mellitus Metabolic Disease). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2021.06.29.21259606\u003c/span\u003e\u003cspan address=\"10.1101/2021.06.29.21259606\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu T, Li K, Herrero P, Georgiou P (2021) Deep Learning for Diabetes: A Systematic Review. IEEE J Biomedical Health Inf 25(7):2744\u0026ndash;2757. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/JBHI.2020.3040225\u003c/span\u003e\u003cspan address=\"10.1109/JBHI.2020.3040225\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZiyadullaev D, Muhamediyeva D, Madazimov K, Madazimov M, Temirov P, Abdukadirov D (2024) Application of ensemble machine learning methods for diabetes diagnosis. \u003cem\u003eBIO Web of Conferences\u003c/em\u003e, \u003cem\u003e121\u003c/em\u003e, 01002. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1051/bioconf/202412101002\u003c/span\u003e\u003cspan address=\"10.1051/bioconf/202412101002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Health Organization (2024), April 12 \u003cem\u003eDiabetes\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/news-room/fact-sheets/detail/diabetes\u003c/span\u003e\u003cspan address=\"https://www.who.int/news-room/fact-sheets/detail/diabetes\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"University of Ibadan","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Diabetes, Deep learning, Two-stage model","lastPublishedDoi":"10.21203/rs.3.rs-9081794/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9081794/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eDiabetes mellitus is a growing global health challenge, particularly in low- and middle-income countries, where delayed diagnosis and inadequate treatment contribute significantly to complications and mortality. Artificial intelligence, particularly deep learning (DL), has emerged as a promising tool for improving diabetes classification and outcome prediction.\u003c/p\u003e\u003ch2\u003eObjective\u003c/h2\u003e \u003cp\u003e This scoping review aimed to map existing evidence on the development and application of one-stage and two-stage deep learning models for the classification and prediction of Type 1 and Type 2 diabetes.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA scoping review was conducted using PubMed and Google Scholar databases, guided by the Population-Concept-Context (PCC) framework and PRISMA-ScR methodology. Studies were included if they applied deep learning models to the classification and/or prediction of diabetes. Data extraction was performed using a structured spreadsheet capturing model type, dataset, features, and performance metrics.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eOut of 750 identified studies, 50 met the inclusion criteria. Convolutional neural network-based architectures were the most common (16; 38%), followed by recurrent neural networks and hybrid models. The majority of studies (43; 86%) used a one-stage deep learning approach integrating classification and prediction into a single step. Only 7 studies (14%) employed a two-stage framework, and none were conducted in the African context. Common datasets included the Pima Indian dataset and the UCI Machine Learning Repository, with limited use of local or clinical datasets.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eDeep learning models demonstrate strong potential for improving diabetes diagnosis and prediction. However, the dominance of one-stage models and the lack of African-based studies highlight critical methodological and geographical gaps. Future research should explore two-stage models tailored to local datasets to enhance clinical relevance and promote global equity in AI-based diabetes care.\u003c/p\u003e","manuscriptTitle":"Scoping Review on Deep Learning Model for Classification and Prediction of Diabetes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-11 10:37:28","doi":"10.21203/rs.3.rs-9081794/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"bf669470-640f-43bb-bac9-3ea69a0b2ca1","owner":[],"postedDate":"March 11th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":64235742,"name":"Endocrinology \u0026 Metabolism"},{"id":64235743,"name":"Epidemiology"},{"id":64235744,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2026-03-11T10:37:28+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-11 10:37:28","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9081794","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9081794","identity":"rs-9081794","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}