{"paper_id":"195b2231-081c-4561-9200-4649b670b576","body_text":"BIBLIOTECA  \n \n \nThis work is licensed under a  \nCreative Commons Attribution-NonCommercial-NoDerivatives  \n4.0 International License. \n       \n \n \n \nDocument downloaded from the institutional repository of the University of \nAlcala: http://ebuah.uah.es/dspace/ \n \nThis is a posprint version of the following published document: \n \nBautista, A., Tardillo, J., Castillo Sequera, J.L. & Wong, L. 2025, “Model for \nendometriosis detection using machine learning algorthms”, in 2025 7th \nInternational Conference on Software Engineering and Computer Science \n(CSECS). \n \nAvailable at https://dx.doi.org/10.1109/CSECS64665.2025.11009460 \n \n© 2025 IEEE. Personal use of this material is permitted. Permission from \nIEEE must be obtained for all other users, including reprinting/republishing \nthis material for advertising or promotional purposes, creating new \ncollective works for resale or redistribution to servers or lists, or reuse of \nany copyrighted components of this work in other works. \n \n(Article begins on next page) \n \n \n\nModel for Endometriosis Detection using Machine \nLearning Algorithms \nAlexis Bautista  \nInformation Systems \nEngineering Program \nUniversidad Peruana de \nCiencias Aplicadas  \nLima, Peru \nu20181a323@upc.edu.pe \nJahir Tardillo \nInformation Systems \nEngineering Program  \nUniversidad Peruana de \nCiencias Aplicadas  \nLima, Peru \nu201713616@upc.edu.pe \nJosé Luis Castillo-Sequera \nDepartment of Computer \nScience \nUniversidad de Alcalá  \nAlcalá de Henares, Spain \njluis.castillo@uah.es \nLenis Wong \nInformation Systems Engineering \nProgram Universidad Peruana de \nCiencias Aplicadas  \nLima, Peru \npcsilewo@upc.edu.pe \nAbstract—Endometriosis is a chronic disease that affects a \nconsiderable percentage of women of reproductive age and is \ncharacterized by the presence of endometrial tissue outside the \nuterine cavity, leading to symptoms such as pelvic pain and \ndysmenorrhea. The aim of this study is to develop a predictive \nmodel for the classification of endometriosis using four Machine \nLearning algorithms: Random Forest, LASSO, SVM, and Naive \nBayes. For this purpose, a dataset from the Global Health Data \nExchange was utilized, consisting of 1,000 cases of patients with \nendometriosis. The methodology included data cleaning and \npreprocessing, as well as the evaluation of each algorithm's \nperformance using four metrics: precision, recall, F1-Score, and \naccuracy. The findings revealed tha t the Random Forest \nalgorithm was the most effective in identifying endometriosis, \noutperforming the other algorithms with a precision of 0.99 for \nthe \"endometriosis\" class and an overall accuracy of 0.98. \nKeywords—endometriosis, machine learning, Random Forest, \nLASSO \nI. INTRODUCTION\nEndometriosis, as defined in [16], is the presence of \nfunctional endometrial tissue (glands and stroma) outside the \nuterine cavity. Additionally, it is a chronic, periodically \nsymptomatic, estrogen-dependent disease that affects between \n10% and 30% of wome n of reproductive age and older. \nAccording to [13], the primary symptom of endometriosis is \npelvic pain, which may occur during vaginal bleeding \n(dysmenorrhea), during sexual intercourse (dyspareunia), or \nindependently of vaginal bleeding (non -menstrual pe lvic \npain). Patients may also experience lower back pain or \nabdominal discomfort. These symptoms can significantly \nimpact a patient’s physical, mental, and social well -being, \nthereby impairing their quality of life. Endoscopic \nSubmucosal Dissection (ESD) and Comprehensive Sexuality \nEducation (CSE) provide clinically relevant evaluations of \nendometriosis symptoms and the disease's impact on patients' \nlives. \nContemporary medicine stands at a critical intersection \nbetween the growing volume of available medical data and the \nneed for more precise and personalized tools for disease \ndiagnosis, treatment, and management. In this context, \nMachine Learning (ML) has emerged as a powerful tool with \nthe potential to radically transform medical practice. \nAccording to [19], ML is used in the medical field to analyze \nlarge datasets, including clinical, genetic, and medical \nimaging data, with the goal of improving diagnosis, treatment, \nand disease management. Beyond endometriosis, ML has also \nproven essential in other medical domains. For instance, [20] \nhighlights its significance in predicting immunotherapy \nresponses in cancer. \nThe integration of Machine Learning (ML) with \nendometriosis research offers significant potential, as ML's \ncapability to process large datasets and detect underlying \npatterns positions it as a key tool for overcoming diagnostic \nchallenges in this gynecological condition. By leveraging ML \nalgorithms, researchers can identify patterns, risk factors, and \nspecific traits associated with endometriosis, providing critical \ninsights for treatment strategies [6]. Additionally, as noted in \n[4] and [18], these technologies improve diagnostic accuracy\nand outcome predictions, fostering the development of more\naccessible non -invasive diagnostic methods. However, as\nhighlighted in [2], the lack of reproducibility in studies\ninvolving microRNAs remains a limitation, underl ining the\nnecessity for consistent and reliable results. Early detection of\nendometriosis is vital, as it enables the application of more\neffective management approaches and enhances patients’\nquality of life [9]. While ML demonstrates the potential for\nmore precise and expedited diagnoses compared to traditional\nmethods, it complements other medical domains that often\nrequire longer and riskier diagnostic pathways for patients.\n With the rapid advancement of technology, this research \nproposes implementing a model for detecting endometriosis \nby applying four Machine Learning algorithms: Random \nForest, Support Vector Machine, LASSO, and Naive Bayes. \nThe implementation follows the CRISP-DM methodology. \nII. RELATED WORKS\nIn the literature, various ML algorithms have been applied \nto address diseases with different purposes, such as analyzing \ntissue samples, identifying genes, and biomarkers. The most \nused algorithms include RF, LASSO, and SVM (see Table 1). \nThe Random Forest (RF) algorithm has demonstrated \nremarkable effectiveness in detecting endometriosis, excelling \nin processing large datasets and improving classification \naccuracy for biomarkers and genetic patterns associated with \nthe disease. For instanc e, studies such as [15] and [17] \nachieved AUC values of 0.721 and 0.939, respectively, by \nemploying RF to classify genes linked to cuproptosis and \nmicroRNAs, underscoring the algorithm’s ability to analyze \ncomplex data. Additionally, research presented in [14] and [8] \nreported AUC values of 0.8226 and 0.895 when identifying \ncritical biomarkers and analyzing genes associated with \nsenescence. Despite variations in study methodologies, such \nas ovarian lesion classification in [3], which achieved an AUC \nof 0.968, RF consistently proves to be a versatile and robust \ntool in detecting gynecological diseases, particularly \nendometriosis.  \nThe LASSO algorithm has demonstrated significant \nefficacy in detecting endometriosis and other gynecological \nconditions, primarily due to its capability to reduce \n\ndimensionality and select relevant features. For example, [8] \nemployed LASSO to identify genes associated with \nendometriosis and senescence, achieving AUC values of \n0.822 during training and 0.895 in validation, highlighting its \nstrong classification performance for endometriosis patients. \nSimilarly, [11] utilized LASSO to select eight key genes \nrelated to M2 macrophages, achieving an AUC ≥ 0.65, \nenabling a detailed analysis of disease severity based on gene \nexpression. Furthermore, [18] demonstrated the al gorithm's \nability to enhance sensitivity and specificity in a non-invasive \ndiagnostic model for endometriosis, achieving an AUC of \n0.80 for the CA125 marker. In addition, [15] applied LASSO \nto mitigate overfitting in their classification model, achieving \nAUC values of 0.781 in training and 0.721 in testing, \nshowcasing its versatility across various applications. \nNotably, [3] utilized LASSO to develop a logistic regression \nmodel for predicting ovarian cancer, achieving an AUC of \n0.946, underscoring its robustness in discriminating between \nbenign and malignant lesions. \nThe Support Vector Machine (SVM) algorithm has proven \nto be highly effective in disease classification and detection, \nparticularly in the medical field, due to its capacity to manage \ncomplex datasets. For instance, [3] utilized SVM to \ndistinguish between b enign and malignant ovarian lesions, \nachieving an AUC of 0.821, which indicates moderate \ndiscriminatory capability. This application was compared \nwith other algorithms, highlighting SVM's ability to identify \npatterns within critical clinical and serological data for precise \ndiagnoses. Similarly, [7] applied SVM for the early diagnosis \nof endometriosis based on self -reported patient data, \nachieving an AUC of 0.87, which demonstrates strong \npredictive performance. While [3] emphasized ovarian lesion \nclassification, [7] showcased SVM's adaptability in \naddressing other clinical challenges such as endometriosis. \nBoth studies emphasize the robustness of SVM in binary \nclassification tasks, although its performance varies \ndepending on the specific medical issue and the characteristics \nof the dataset being analyzed. \nTABLA 1.  \nAlgorithm Purpose AUC Ref. \nRF Distinguish endometrial tissue \nsamples \n0.85 [8] \nRF Identify genes based on the \nsamples \n0.78 [15] \nRF Classify the presence or \nabsence of endometriosis \n0.94 [17] \nRF Identify genes associated with \nendometriosis \n0.78 [14] \nRF Identify key genes related to \nM2 macrophages in \nendometriosis \n0.65 [11] \nRF Build predictive models to \nclassify ovarian lesions. \n0.96 [3] \nLASSO Distinguish features among \ngenes expressed with \nendometriosis. \n0.85 [8] \nLASSO Identify key genes related to \nM2 macrophages in \nendometriosis. \nAUC > 0.65 [11] \nLASSO Classify patients based on their \nsymptoms and self -reported \ncharacteristics. \nAUC  [7] \nLASSO Classify segmented images for \ntumor recognition. \nAUC > 0.85 [12] \nLASSO Distinguish characteristics \namong genes expressed with \nAUC > 0.85 [8] \nendometriosis. \nLASSO Identify key genes associated \nwith M2 macrophages in \nendometriosis. \nAUC > 0.65 [11] \nLASSO Improve diagnostic accuracy. AUC > 0.8 [18] \nLASSO Identify genes that aid in the \nprediction of endometriosis. \nAUC > 0.78 [15] \nLASSO Filter the most relevant \npredictors among serological \nbiomarkers. \nAUC > 0.96 [3] \nSVM Classification model to predict \nthe presence of ovarian lesions. \nAUC > 0.96 [3] \nSVM Classify data within the context \nof medical decision support \nsystems. \nAUC > 0.88 [1] \nSVM  Identify genes serving as \ndiagnostic markers for \npulmonary arterial \nhypertension. \nAUC > 0.94 [5] \nIII. PROPOSED MODEL\nFor this reason, none of these studies focus on the \nidentification of patient symptoms. Our proposal emphasizes \nthis aspect because symptoms are key clinical indicators for \nthe early detection of endometriosis, a condition often \ndiagnosed years after its onset due to its variability and \ncomplex presentation. By focusing on symptoms reported by \npatients, we aim to develop a prediction system that not only \nfacilitates faster and non-invasive diagnoses but also provides \na practical tool for physicians in clinical settings. This \napproach allows for a personalize d strategy tailored to the \nindividual characteristics of each patient, thereby improving \nthe accuracy and effectiveness of treatment from early stages.\nFig. 1. Conceptual Model of the Proposed Approach. \nA. Data selection\nFor this study, a dataset extracted from the Global Health\nData Exchange (GHDx) site [9] will be used, containing 250 \ncases of patients with endometriosis from the year 2020. Each \ncase includes 15 features related to symptoms and other \nclinically relevant variables for diagnosing the disease. \nB. Procesamiento de Data\nData preprocessing techniques are critical to ensuring data\nquality and enhancing the performance of Machine Learning \nmodels. For this study, the data preprocessing phase involved \nseveral steps: data cleaning, feature selection, and variable \nencoding for numerical representation using Python functions \nlike ‘.map’. For instance, to encode the variable \"Infertility,\" \nwhich initially contains the values 'Yes' and 'No,' the ‘.map’ \nfunction was applied to convert these values into numerical \nvariables such as '1' for 'Yes' and '0' for 'No.' Table II presents \nthe nine features selected. \nTABLE II. DATASET CHARACTERISTICS \nID Characteristics Description \nF01 Age Patient's age \nF02 Ethnicity Human community \nF03 Days of Menstruation Duration of menstruation in \ndays \nF04 Pain Presence of pain \n\n\nF05 Increased Bleeding Increased menstrual \nbleeding \nF06 Prolonged \nMenstruation \nExtended menstrual cycle \ndays \nF07 Infertility Difficulty conceiving \nF08 Dyspareunia Genital pain during sexual \nintercourse \nF09 Dysmenorrhea Lower abdominal pain \nC. Algorithm Training\nFor this study, the three most used  algorithms in the\nliterature were employed: Random Forest, LASSO, and SVM \n(Support Vector Machine). \nIn Figure 2, the steps for training the Random Forest \nalgorithm to predict endometriosis diagnosis in Python are \nillustrated \nThe flowchart illustrates the process of using a machine \nlearning model for diagnosing endometriosis. Initially, patient \ndata and reported symptoms are provided as input for analysis. \nIn Step 1, the dataset is divided into training and testing \nsubsets using the train_test_split function, allocating 70% for \ntraining and 30% for testing. Step 2 involves training the \nRandom Forest algorithm with the fit function on the training \ndata (X_train and Y_train), enabling it to learn patterns \nassociated with endometri osis symptoms. In Step 3, the \ntrained model predicts outcomes for the test data (X_test), \nwith results stored in predictions. A confusion matrix \nevaluates the model's performance by comparing correct and \nincorrect predictions. Step 4 calculates accuracy us ing the \naccuracy_score function to determine the model's \neffectiveness in classifying cases. Additionally, the model's \nreliability is validated using real clinical data (x_real_data) \nunseen during training. Finally, the output presents the \ndiagnosis, indic ating whether the patient is likely to have \nendometriosis. This methodology was consistently applied to \nother algorithms evaluated in the study. \nFig. 2. Steps for Training the Random Forest Model in Python. \nThis process was carried out in a Python -based \ndevelopment environment, specifically using PyCharm. \nPopular libraries such as pandas, scikit -learn, and matplotlib \nwere employed for data manipulation, model construction, \nand result visualization. \nIn the correlation analysis of the endometriosis dataset \nfeatures, a correlation matrix was utilized to identify \nrelationships between different variables and the diagnosis of \nendometriosis. This matrix aids in visualizing the \ncompatibility of each feature with the desired outcome, in this \ncase, the \"Endometriosis Diagnosis.\" It was observed that \ncertain variables, such as F05 (Increased bleeding) and F09 \n(Dysmenorrhea), show significant correlation with the \ndiagnosis, which is critical for optimizing the p redictive \nmodel. This approach streamlines the feature selection process \nand enhances the model's accuracy by focusing on the most \nrelevant variables for prediction. \nFig. 3. Features of the Correlation Matrix and Dataset. \nD. Evaluation\nAfter the classification process, the results are compared\nwith the actual diagnoses from the dataset to generate four key \nvariables: \"True Positives\" (TP), \"False Positives\" (FP), \n\"False Negatives\" (FN), and \"True Negatives\" (TN). These \nvariables are essential for calculating metrics that evaluate the \nmodel's performance, as explained in Table IV. \nTABLE IV. DESCRIPTION OF DATASET VARIABLES \nVariable Description \n TP Cases where the model correctly predicted that a patient has \nendometriosis, and she actually has the disease. \nFP Cases where the model predicted that a patient has \nendometriosis, but she actually does not have the disease. \nFN Cases where the model did not predict endometriosis \n(negative prediction), but the patient does have the disease. \nTN Cases where the model correctly predicted that a patient \ndoes not have endometriosis, and she actually does not have \nthe disease. \nThe analysis includes the following metrics: Precision (Eq. \n1), Recall (Eq. 2), F1 Score (Eq. 3), and Accuracy (Eq. 4), \nwhich provide a comprehensive view of how the model \npredicts endometriosis in patients. These metrics are presented \nin the following equations: \n𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 =  𝑇𝑃\n𝑇𝑃 + 𝐹𝑃\n(1) \n\n\n𝑅𝑒𝑐𝑎𝑙𝑙 =  𝑇𝑃\n𝑇𝑃 + 𝐹𝑁\n(2) \n𝐹1 𝑠𝑐𝑜𝑟𝑒 = 𝑇𝑁\n𝑇𝑁 + 𝐹𝑃\n(3) \n𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 = 𝑇𝑃 + 𝑇𝑁\n𝑇𝑃 + 𝑇𝑁 + 𝐹𝑃 + 𝐹𝑁\n(4) \nIV. RESULTS AND DISCUSSION\nFigure 4 illustrates the confusion matrices for the three \nalgorithms: RF (Figure 4a), LASSO (Figure 4b), SVM (Figure \n4c), and Naive Bayes (Figure 4d). Additionally, Table 5 \nsummarizes the number of correct and incorrect predictions \nfor each trained model. The results of the confusion matrix \nmetrics for the four algorithms applied to endomet riosis \ndetection are as follows: Random Forest achieved the highest \nprecision, with 237 correct predictions of endometriosis and a \nlow error rate (only 2 errors in each category). LASSO and \nSVM also delivered strong results, although SVM recorded \nmore erro rs (11) in the \"No Endometriosis\" category. In \ncontrast, Naive Bayes exhibited the highest number of errors \nacross both categories. \n(a) (b) \n(c) (d) \nFig. 4. Confusion Matrices for RF (a), LASSO (b), SVM (c), and Naive \nBayes (d). \nTABLE V. CONFUSION MATRIX METRICS FOR THE DATASET \nAlgorithm Feature Correct \nPrediction \nIncorrect \nPrediction \nTotal \nRandom \nForest \n0=’No \nEndometriosis’ \n60 2 62 \n1=’ \nEndometriosis’ \n237 2 239 \nLASSO 0=’No \nEndometriosis’ \n60 2 62 \n1=’ \nEndometriosis’ \n225 14 239 \nSVM 0=’No \nEndometriosis’ \n51 11 62 \n1=’ \nEndometriosis’ \n238 1 239 \nNaive \nBayes \n0=’No \nEndometriosis’ \n56 6 62 \n1=’ \nEndometriosis’ \n214 25 239 \nFigure 5 presents the results of the algorithms based on the \narea under the curve (AUC) metrics, where a value closer to 1 \nindicates a more effective classifier. The results reveal that the \nSVM algorithm (Figure 5c) achieved the highest classification \nperformance with an AUC of 0.99, followed by Random \nForest (Figure 5b). \n(a) (b) \n(c) (d) \nFig. 5. ROC Curve for RF (a), LASSO (b), SVM (c), and Naive Bayes (d). \nTable VI displays the training metrics for the four \nalgorithms, highlighting that the Random Forest algorithm \ndemonstrated strong performance with high precision, recall, \nand F1 score for both \"no endometriosis\" (0) and \n\"endometriosis\" (1) cases. This indi cates significant \neffectiveness in classifying both types of samples. The \nLASSO algorithm also delivered satisfactory results, showing \nbetter precision in classifying the \"endometriosis\" (1) class. \nSVM exhibited higher precision for the classification of t he \n\"endometriosis\" (1) class. Finally, the Naive Bayes model \nshowed a marked difference in precision between classes, \nwith the \"endometriosis\" class achieving the highest accuracy. \nTABLE VI TRAINING METRICS RESULTS FOR THE DATASET \nAlgorithm Feature Precision Recall F1 \nScore \nACC \nRandom \nForest \n0 0.96 0.96 0.96 0.98 \n1 0.99 0.98 0.99 \nLASSO 0 0.81 0.96 0.88 0.96 \n1 0.94 0.99 0.96 \nSVM 0 0.98 0.82 0.89 0.96 \n1 0.95 0.98 0.97 \nNaive Bayes 0 0.70 0.90 0.78 0.90 \n1 0.97 0.89 0.89 \n\n\nV. CONCLUSION\nIn this study, a model for the detection of endometriosis \nwas proposed, focusing on symptoms reported by patients and \napplying four machine learning algorithms: Random Forest, \nLASSO, SVM, and Naive Bayes. The model was developed \nusing a dataset extracted from the Global Health Data \nExchange (GHDx), comprising 1,000 cases of patients with \nendometriosis. These algorithms were applied and fine-tuned, \nwith 70% of the data used for training and the remaining 30% \nreserved for evaluation. Key metrics such as precision, recall, \naccuracy, and F1 -Score were employed to assess the \nperformance of each algorithm, providing a comprehensive \nanalysis of their effectiveness. The results demonstrated that \nRandom Forest was the most effective algorithm, followed by \nLASSO and SVM, while Naive Bayes exhibited lower \nperformance. Collectively, these findings highlight the \nfeasibility of the proposed approach to enhance early detection \nof endometriosis and its potential impact on clinical practice. \nAs a future endeavor, it is suggested to develop a tool \nutilizing the Random Forest algorithm to provide healthcare \nprofessionals with a practical and personalized system, \ntransforming clinical practice and delivering significant \nbenefits to a wide range of women of reproductive age. \nACKNOWLEDGMENT \nWe extend our gratitude to the Research Department of the \nUniversidad Peruana de Ciencias Aplicadas for their support \nin the execution of this project. \nREFERENCES \n[1] A. Awaysheh, J. Wilcke, F. Elvinger, L. Rees, W. Fan, y K. L. \nZimmerman, \"Review of medical decision support and machine\nlearning methods,\" Journal of Veterinary Diagnostic Investigation , \nvol. 31, no. 1, pp. 1-14, 2019. doi: 10.1177/1040638718801234.\n[2] A. Leonova, V. E. Turpin, S. K. Agarwal, M. Leonardi, y W. G. Foster, \n\"A critical appraisal of the circulating levels of differentially expressed \nmicroRNA in endometriosis,\" Biology of Reproduction, vol. 105, no. \n5, pp. 1075-1085, 2021. doi: 10.1093/biolre/ioab134.\n[3] B. Jing, Y. Zhang, Y. Liu, Y. Wang, Y. Chen, Y. Zhang, Y. Li, Y. \nZhang, Y. Wang, and Y. Liu, \"Development of prediction model to\nestimate future risk of ovarian lesions,\" Preventive Medicine Reports, \nvol. 35, p. 102296, 2023. doi: 10.1016/j.pmedr.2023.102296.\n[4] B. Sivajohan, M. Elgendi, C. Menon, C. Allaire, P. Yong, and M. A.\nBedaiwy, \"Machine learning algorithms as new screening approach for \npatients with endometriosis,\" Sci. Rep., vol. 12, p. 639, 2022. DOI:\n10.1038/s41746-022-00638-1 \n[5] C. Jiang, Y. Jiang, \"Lasso Algorithm and Support Vector Machine \nStrategy to Screen Pulmonary Arterial Hypertension Gene,\" Scottish \nMedical Journal , vol. 68, no. 1, pp. 24 -30, 2022. doi: \n10.1177/00369330211064645. \n[6] C. Xie, Z. Yin, and Y. Liu, \"Analysis of characteristic genes and\nceRNA regulation mechanism of endometriosis based on full \ntranscriptional sequencing,\" Frontiers in Genetics, vol. 13, p. 902329, \n2022. doi: 10.3389/fgene.2022.902329.\n[7] D. B. Balogh, G. Hudelist, D. Bliznuks, J. Raghothama, C. M. Becker,\nR. Horace, H. Krentel, A. W. Horne, N. Bourdel, G. Marki, C. \nTomassetti, U. B. Kirk, N. Acs, y A. Bokor, \"FEMaLe: The use of \nmachine learning for early diagnosis of endometriosis based on patient \nself-reported data—Study protocol of a multicenter trial,\" PLoS ONE, \nvol. 19, no. 5, p. e0300186, 2024. doi: 10.1371/journal.pone.0300186. \n[8] D. Zou, L. Zhang, Y. Xu, K. Wang, and J. Zhang, \"Identification of \nsenescence hub genes and their association with endometriosis,\" \nFrontiers in Pharmacology , vol. 14, p. 1259467, 2023. doi:\n10.3389/fphar.2023.1259467. \n[9] F. G. Martire, L. Lazzeri, F. Conway, T. Siciliano, A. Pietropolli, E.\nPiccione, E. Solima, G. Centini, E. Zupi, and C. Exacoustos,\n\"Adolescence and endometriosis: symptoms, ultrasound signs and\nearly diagnosis,\" Fertility and Sterility, vol. 114, no. 5, pp. 1049-1057,\nNov. 2020. DOI: 10.1016/j.fertnstert.2020.06.041.\n[10] Global Health Data Exchange, “Global Health Data Exchange,”\n[Online]. Available: https://ghdx.healthdata.org/. \n[11] H. Ding, H. Xu, T. Zhang, and C. Shi, \"Identification and validation of \nM2 macrophage -related genes in endometriosis,\" Heliyon, vol. 9, \ne22258, 2023. doi: 10.1016/j.heliyon.2023.e22258.\n[12] H. I. H. Inbarani, A. T. Azar, and J. G., \"Leukemia Image Segmentation \nUsing a Hybrid Histogram -Based Soft Covering Rough K -Means\nClustering Algorithm,\" Electronics, vol. 9, no. 188, pp. 1-22, Jan. 2020. \ndoi: 10.3390/electronics9010188. \n[13] H. Kitchen, C. Seitz, A. Trigg, N. Aldhouse, T. Willgoss, H. Schmitz, \nA. Gater, C. Gerlinger, and C. Haberland, \"Patients’ and clinicians’ \nperspectives on item importance, scoring, and clinically meaningful \ndifferences for the Endometriosis Symptom Diary ( ESD) and \nEndometriosis Impact Scale (EIS),\" Health Qual. Life Outcomes, vol.\n19, no. 7, 2021. DOI: 10.1186/s12955 -020-01579-7. [Online]. \nAvailable: https://doi.org/10.1186/s12955-020-01579-7 \n[14] J. She, D. Su, R. Diao, y L. Wang, \"A Joint Model of Random Forest \nand Artificial Neural Network for the Diagnosis of Endometriosis,\"\nFrontiers in Genetics , vol. 13, p. 848116, 2022. doi: \n10.3389/fgene.2022.848116. \n[15] J. Wei, B. Huang, W. Liu, Q. Zhang, Y. Nong, W. Wang, X. Liang, Q. \nLi, Y. Xie, T. Peng, \"Identification of cuproptosis-related differentially\nexpressed genes and their clinical significance in endometriosis,\"\nReproductive Sciences , vol. 30, no. 11, pp. 1841 -1853, 2023. doi:\n10.1007/s43032-022-01130-7. \n[16] M. Anantha Satyanarayana, VV Sushma, K Praveen Kumar, V Sudheer \nKumar, K Rama Murthy, and T Asha, \"Endometriosis: A \nClinicopathological Study of 29 Cases,\" Journal of Clinical and \nDiagnostic Research, vol. 12, no. 5, pp. EC05-EC09, May 2018. DOI:\n10.7860/JCDR/2018/35204.11460. [Online]. Available:\nhttps://doi.org/10.7860/JCDR/2018/35204.11460\n[17] M. F. Moustafa, A. M. Al-Hendy, A. M. El-Shal, A. A. El-Badawy, A. \nA. El-Sayed, A. A. El-Sharkawy, A. A. El-Masry, A. A. El-Sayed, and \nA. A. El-Masry, \"Endometriosis diagnosis by microRNA biomarkers,\" \nAmerican Journal of Obstetrics & Gynecology , vol. 223, no. 4, pp. \n557.e1-557.e12, 2020. doi: 10.1016/j.ajog.2020.06.014.\n[18] M. Szubert, A. Rycerz, and J. R. Wilczyński, \"How to Improve Non -\nInvasive Diagnosis of Endometriosis with Advanced Statistical\nMethods,\" Medicina, vol. 59, no. 3, p. 499, 2023. doi:\n10.3390/medicina59030499. \n[19] R. Khamisy -Farah, L. B. Furstenau, J. D. Kong, J. Wu, and N. L.\nBragazzi, \"Gynecology Meets Big Data in the Disruptive Innovation\nMedical Era: State -of-Art and Future Prospects,\" in International \nJournal of Environmental Research and Public Health, vol. 18, p. 5058, \nMay 2021. [Online]. Available: \nhttps://doi.org/10.3390/ijerph18105058. \n[20] W. Jin, Q. Yang, H. Chi, K. Wei, P. Zhang, G. Zhao, S. Chen, Z. Xia, \nand X. Li, \"Ensemble deep learning enhanced with self -attention for \npredicting immunotherapeutic responses to cancers,\" Front. Immunol.,\nvol. 13, pp. 1025330, Dec. 2022. [Online]. Availab le:\nhttps://doi.org/10.3389/fimmu.2022.1025330.","source_license":"CC0","license_restricted":false}