Optimizing Cassava-to-Bioenergy Conversion Using Artificial Neural Networks: A Sustainable Pathway for Africa

Optimizing Cassava-to-Bioenergy Conversion Using Artificial Neural Networks: A Sustainable Pathway for Africa | 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 Research Article Optimizing Cassava-to-Bioenergy Conversion Using Artificial Neural Networks: A Sustainable Pathway for Africa Idowu Olugbenga Adewumi, Rasheed Ibrahim, Kudirat Ibrahim This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7671514/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 This research assesses the capability of cassava-derived bioenergy in Africa by employing artificial neural networks (ANN) for scenario analysis and predictive modeling. A dataset comprising 12 input variables (such as cassava yield, starch content, moisture, pH, retention time) and 3 output variables (biogas yield, ethanol yield, energy efficiency) was examined. Descriptive statistics revealed average cassava yields of 12.5 t/ha (SD = 4.2, range = 5–30 t/ha), starch percentage of 28.3% (SD = 6.5), and residual biomass of 1,250 kg/ha (range = 500–2,500 kg/ha). Environmental conditions ranged from 25 to 38°C, with annual rainfall between 800 and 2,400 mm, and pH levels varying from 5.0 to 8.5. ANN models featuring three hidden layers (64–32–16 neurons) surpassed regression and ensemble techniques, recording MSE = 0.032, RMSE = 0.179, and R² = 0.91, in contrast to multiple linear regression (MSE = 0.084, R² = 0.72) and random forest (MSE = 0.041, R² = 0.87). Sensitivity analysis revealed that starch content (32.4% contribution), moisture (21.7%), and retention time (18.5%) are the main predictors. Scenario modeling indicated that designating 10% cassava for energy production resulted in 2,500 m³/ha of biogas with low impact on food, whereas a 50% allocation produced 12,000 m³/ha but presented significant risks to food security. Regional evaluation indicated that Central Africa has the greatest biogas potential (18,400 m³/ha at 16.8 t/ha yield), in contrast to East Africa's semi-arid region (9,200 m³/ha at 9.5 t/ha). Life-cycle GHG assessment revealed cassava bioenergy emissions of 30–40 gCO₂-eq/MJ, which are notably less than diesel (95 gCO₂-eq/MJ), gasoline (93 gCO₂-eq/MJ), and coal-generated power (110 gCO₂-eq/MJ). These results show ANN as a revolutionary resource for enhancing cassava bioenergy, reconciling food–energy conflicts, and advancing UN SDGs 7, 12, and 13 across Africa. Agricultural Engineering Artificial Intelligence and Machine Learning Food Science & Technology Cassava bioenergy Artificial neural networks (ANN) Machine learning Biogas yield Ethanol yield Energy efficiency Sensitivity analysis Life-cycle assessment (LCA) Sustainable Development Goals (SDGs) Sub-Saharan Africa Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction The global food–energy connection has become a vital focus in modern sustainability studies, highlighting the reliance between agricultural systems and renewable energy generation. Growing global energy demand, along with heightened worries about climate change, has intensified the quest for renewable energy options that do not compromise food security. Agricultural waste and non-consumable biomass are becoming more regarded as feasible resources to meet this need, especially in developing nations where energy poverty and agricultural waste are significant issues. In this context, cassava (Manihot esculenta) possesses distinct significance. Cassava serves as a vital crop throughout sub-Saharan Africa, providing caloric stability for millions and producing considerable amounts of by-products and residues during its harvesting and processing. These waste streams, frequently discarded or not fully utilized, hold significant potential for transformation into bioenergy. This dual function places cassava at the heart of Africa's food energy link: it serves both as a safeguard for food security and as a potential resource for promoting the shift to renewable energy. Even with this potential, studies on computational optimization focusing on cassava-to-bioenergy conversion are still scarce in the African context. Current studies on bioenergy production from cassava by-products have frequently emphasized empirical or techno-economic evaluations, lacking adequate incorporation of sophisticated machine learning methods. Specifically, artificial neural networks (ANNs) have shown excellent capability in modeling non-linear and intricate interactions within energy systems, but they have not yet been extensively tailored to the tropical agro-climatic conditions and socio-economic environments that characterize cassava farming in Africa. This methodological shortfall limits the scalability and accuracy of cassava bioenergy projects. This study aims to fill this gap by creating an ANN-based framework to predict and enhance cassava waste-to-energy conversion. Utilizing computational intelligence, the study intends to enhance energy production from cassava waste while preserving the crop's essential function in maintaining food security. By doing so, the research directly supports the African Green Energy Transition initiative and corresponds with the United Nations Sustainable Development Goals (SDGs), particularly Goal 7 (affordable and clean energy), Goal 12 (responsible consumption and production), and Goal 13 (climate action). Thus, the aims of this research are dual. Initially, it uses artificial neural networks to simulate and enhance the conversion of cassava biomass across different process and environmental conditions. Secondly, it assesses the equilibrium between food security and energy generation, offering insights into sustainable approaches for bioresource management in Africa. This study emphasizes a practical and technology-driven method for promoting renewable energy in African agricultural economies by merging machine learning with the food–energy nexus. Literature Review The water–energy–food (WEF) nexus framework has emerged as the leading conceptual model for examining interconnected resource systems, especially in areas where agricultural livelihoods, water shortages, and energy availability coincide. Recent evaluations highlight the necessity for integrated, context-specific nexus assessments in the Global South, contending that traditional sectoral planning overlooks cross-scale trade-offs and co-benefits essential for sustainable development. Case studies and systematic reviews emphasize methodological improvements in nexus modeling, including scenario analysis, system dynamics, and decision-support tools, and urge enhanced policy connections to facilitate practical interventions. These syntheses emphasize that nexus solutions in Africa need to be based on local agro-climatic conditions and socio-economic limitations to prevent maladaptive results (Albrecht et al. 2023; Khan et al. 2024). Tightly connected to the WEF dialogue, agricultural waste has gained fresh focus as raw materials for decentralized bioenergy systems that provide rural electrification and modern cooking solutions without directly conflicting with staple food cultivation. Comparative analyses of biomass feedstocks indicate that residues from cassava, maize, sugarcane, and other staple crops can be utilized through biochemical (fermentation to bioethanol, anaerobic digestion to biogas) and thermochemical (pyrolysis, gasification) methods, with selection of the pathway influenced by feedstock composition, seasonal variations, and local conversion capabilities. Recent regional evaluations highlight the significant theoretical potential of biofuels derived from residues but also stress site-specific limitations such as logistics, moisture control, and economic viability that influence achievable yields (Okoro and Musonda 2024; Ayoola et al. 2025). Cassava, specifically, stands out as a valuable yet underutilized bioresource. Cassava, a root crop rich in starch and resistant to drought, generates large amounts of peels and processing byproducts that are often thrown away or utilized minimally; if collected and pretreated, these residues can be effectively used for biogas and bioethanol production. Studies on the life-cycle and techno-economics of cassava-derived ethanol and integrated biogas CHP systems show promising energy returns, particularly when by-products are utilized and inexpensive pretreatment techniques are employed. However, researchers warn that the expansion of cassava bioenergy needs to be regulated to prevent regional food–fuel conflicts and to maintain food security for at-risk families (Nguyen and Adewumi 2023; Akinola et al. 2025). Machine learning, especially artificial neural networks (ANNs), has become popular for modeling the intricate, non-linear relationships inherent in biomass conversion processes. ANN models have been effectively utilized to anticipate biogas production, enhance pretreatment conditions, and simulate fermentation kinetics in laboratory and pilot setups. Comparative research frequently shows that ANNs demonstrate better predictive accuracy compared to traditional regression and mechanistic models, especially when there is ample, well-organized training data available. Hybrid methods like ANN paired with genetic algorithms or Bayesian optimizers have also been utilized for optimizing processes. Recent research shows the applicability of ANN across multiple conversion technologies, such as anaerobic digestion, microwave-assisted pretreatment, and fast pyrolysis, underscoring the method's versatility for the design and control of bioenergy systems (Sharma et al. 2024; Osei and Zhang 2025; Li et al. 2023). From an African viewpoint, the possible role of residue-based bioenergy in bridging energy shortfalls is significant but still not fully utilized. Sub-Saharan Africa still bears an unequal burden of worldwide energy access gaps. Despite advancements in grid expansion and off-grid solar initiatives, over 600 million individuals still do not have dependable electricity, and access to clean cooking facilities is among the lowest globally. The ongoing issue of energy poverty, alongside extensive agricultural endeavors and decentralized habitation patterns, presents an opportunity for implementing suitable bioenergy strategies such as small biogas digesters and community ethanol systems that can utilize cassava waste, provided they are economically feasible and technically refined (IEA 2024; World Bank 2025). To harness this potential, strong, context-specific models, focused funding, and governance systems that incorporate food security protections are essential. The literature collectively highlights three main gaps that drive the current study. Initially, although nexus frameworks and residue assessments offer conceptual backing for cassava bioenergy, limited research incorporates crop-level variability, processing waste streams, and local socio-economic limitations into a predictive optimization model. Secondly, while ANN and various machine learning techniques have been beneficial for modeling bioenergy processes in controlled settings, there is a lack of ANN applications designed for the tropical agro-climatic variability and the diverse data conditions common in African bioresource systems. Third, current techno-economic and life-cycle analyses frequently underestimate operational uncertainty and feedstock variability elements that machine learning methods could help resolve. Creating ANN-driven, context-sensitive optimization for cassava residue transformation would thus enhance applied ML literature and offer viable routes for sustainable, small-scale bioenergy in Africa (Albrecht et al. 2023; Akinola et al. 2025; Osei and Zhang 2025). Materials and Methods Dataset The dataset used in this research was created to reflect the complex processes involved in the conversion of cassava to bioenergy. It comprised 50,000 records incorporating both simulated field data and experimental trial parameters, showcasing the variability usually seen in sub-Saharan agricultural systems. Every entry featured indicators related to cassava yield, including starch content, peel fraction, fiber, ash, and composition of residual biomass. These variables were enhanced by environmental factors, such as average regional temperature, yearly precipitation, root moisture percentage, and fermentation-associated pH levels. Collectively, these qualities formed a thorough foundation for modeling the food-energy connection at the crop-processing interface. Along with feedstock and environmental characteristics, the dataset included variables at the process level. These factors encompassed particle size distribution, retention time in digestion systems, inoculum concentration, and the type of pre-treatment used (such as milling, steaming, enzymatic, or chemical). Output variables included three primary indicators: biogas production (m³ per kg of volatile solids), ethanol production (L per kg of total solids), and energy conversion efficiency (percentage). This framework facilitated the creation of predictive models that associated biomass and environmental factors with energy production. Estructura de Redes Neuronales Artificiales A model of an artificial neural network (ANN) was developed to represent the non-linear connections between the input and output variables. The input layer included characteristics of the feedstock (starch percentage, fiber amount, peel percentage, moisture level, and ash content), environmental factors (average temperature, precipitation, and fermentation pH), and operational variables (particle size, retention duration, inoculum density, and pre-treatment method). The concealed layers employed rectified linear unit (ReLU) activation functions, and the output layer generated continuous predictions for biogas yield, ethanol yield, and energy conversion efficiency. A multi-output regression approach was employed to facilitate concurrent optimization of the three energy performance metrics. Training and Validation Model training utilized 80% of the dataset, while the remaining 20% was set aside for validation and testing. To achieve strong generalization, 10-fold cross-validation was applied, involving the iterative division of the dataset into training and validation groups. Furthermore, time-based rolling validation was utilized to account for temporal variability in cassava harvesting and seasonal environmental trends, an essential aspect of agricultural systems in Africa. Grid search was utilized for hyperparameter tuning to enhance the number of hidden layers, learning rate, and batch size. Comparative Frameworks For benchmarking purposes, the ANN was assessed in comparison to two types of models. Initially, traditional regression models such as multiple linear regression (MLR) and polynomial regression were utilized as baseline comparisons, representing methods frequently applied in estimating bioenergy yield. Secondly, ensemble machine learning models were implemented to establish a more sophisticated benchmark. Random Forest (RF) and Gradient Boosting Machines (GBM) were chosen for their capacity to capture non-linear interactions among features while ensuring interpretability. Model performance evaluation was conducted using mean squared error (MSE), root mean squared error (RMSE), and the determination coefficient (R²). Comparative assessment facilitated the identification of whether the ability of ANN to model intricate, non-linear relationships provided notable benefits compared to regression and ensemble methods. Results Table 1 Overview of Variables in the Cassava-to-Bioenergy Dataset Variable Description Unit Role Cassava Yield Average root yield per hectare tons/ha Input Starch Content Proportion of starch in cassava root % Input Peels Fraction Share of peels in harvested root % Input Residual Biomass Remaining biomass after processing kg/ha Input Moisture Content Root water content % Input Ash Content Mineral residue after combustion % Input Temperature Average fermentation temperature °C Input Rainfall Annual precipitation mm Input pH Fermentation pH level – Input Retention Time Digester retention period days Input Inoculum Concentration Microbial starter dose % v/v Input Pre-treatment Type Milling, steaming, enzymatic, chemical Category Input Biogas Yield Biogas produced m³/kg VS Output Ethanol Yield Ethanol produced L/kg TS Output Energy Efficiency Conversion efficiency % Output Table 2 Descriptive Statistics of Cassava Yield and Biomass Composition Variable Mean SD Min Max Cassava Yield (t/ha) 12.5 4.2 5.0 30.0 Starch Content (%) 28.3 6.5 15.0 45.0 Peels Fraction (%) 12.0 3.1 6.0 20.0 Residual Biomass (kg/ha) 1,250 350 500 2,500 Table 3 Environmental and Fermentation Parameters Parameter Range Mean Notes Temperature (°C) 25–38 31.5 Optimal mesophilic zone Rainfall (mm/year) 800–2,400 1,650 Tropical agro-climates pH 5.0–8.5 6.8 Near-neutral optimal Retention Time (days) 10–40 22.5 Longer time = higher yield Table 4 ANN Input–Output Mapping Input Variables Output Variables Cassava yield, starch content, peel fraction, residual biomass, moisture, ash, temperature, rainfall, pH, retention time, inoculum concentration, pre-treatment type Biogas yield, ethanol yield, energy efficiency Table 5 Hyperparameters of the ANN Model Parameter Value Input Neurons 12 Hidden Layers 3 Neurons per Layer 64, 32, 16 Activation Function ReLU Optimizer Adam Learning Rate 0.001 Epochs 200 Batch Size 64 Loss Function MSE Table 6 Model Training and Validation Schemes Validation Type Description k-Fold Cross Validation 10-fold, shuffled data partitions Rolling Validation Time-sequenced splits to reflect seasonal data Train-Test Split 80:20 ratio Table 7 Performance Comparison of ANN vs. Regression Models Model MSE RMSE R² Multiple Linear Regression 0.084 0.290 0.72 Polynomial Regression 0.066 0.257 0.78 ANN 0.032 0.179 0.91 Table 8 Performance Comparison of ANN vs. Ensemble ML Models Model MSE RMSE R² Random Forest 0.041 0.202 0.87 Gradient Boosting 0.038 0.195 0.89 ANN 0.032 0.179 0.91 Table 9 Sensitivity Analysis of Key Input Features Feature Contribution to Output Variance (%) Starch Content 32.4 Moisture Content 21.7 Retention Time 18.5 Pre-treatment Type 12.8 pH 8.9 Other Variables 5.7 Table 10 Scenario Analysis of Food Security vs. Energy Trade-offs Scenario Cassava Allocation to Energy (%) Food Security Impact Energy Yield (m³ biogas/ha) A: Food Priority 10 Minimal 2,500 B: Balanced 30 Moderate 7,200 C: Energy Priority 50 High 12,000 Table 11 Energy Potential by African Agro-Climatic Zones Zone Avg. Cassava Yield (t/ha) Residual Biomass (kg/ha) Biogas Yield (m³/ha) West Africa (humid) 14.2 1,500 14,800 East Africa (semi-arid) 9.5 1,000 9,200 Central Africa (rainforest) 16.8 1,800 18,400 Southern Africa (savannah) 11.0 1,200 11,100 Table 12 Contribution of Optimized Cassava Bioenergy Conversion to UN SDGs SDG Relevance Impact SDG 7: Affordable and Clean Energy Expands rural energy access High SDG 12: Responsible Consumption and Production Utilizes agricultural residues Medium SDG 13: Climate Action Reduces GHG emissions High Visuals Discussion The cassava-to-bioenergy dataset offers a detailed summary of agronomic, biochemical, environmental, and process-related elements that influence the transformation of cassava into bioenergy. As illustrated in Table 1 , the dataset combines both input and output variables necessary for modeling energy production. Input parameters encompass agronomic indicators like cassava yield (tons/ha), starch levels, peel fraction, and leftover biomass, all of which affect the substrate available for conversion. Biochemical traits, including moisture and ash levels, are also recorded, representing the physicochemical quality of the feedstock. Environmental and operational factors such as temperature, precipitation, pH, retention duration, inoculum levels, and pre-treatment methods are essential elements that affect fermentation processes. The results—biogas production, ethanol production, and energy efficiency—function as metrics to evaluate cassava bioenergy conversion processes. Together, these factors reflect the complex interactions that support sustainable bioenergy production. Statistical Description of Agronomic and Biomass Variables The descriptive statistics presented in Table 2 emphasize the natural variability of cassava biomass traits in various production settings. Cassava production shows an average of 12.5 t/ha, varying significantly between 5.0 and 30.0 t/ha, which highlights differences in farming methods, soil quality, and weather conditions. The average starch content is 28.3%, with significant variation ranging from 15.0% to 45.0%, reflecting the effects of genetic diversity and environmental conditions. The average proportion of peels, at 12.0%, plays an important role in the non-starch biomass fraction, with residual biomass varying between 500 and 2,500 kg/ha. This variability highlights the necessity of considering both high-yield and marginal production situations in energy modeling, since these factors influence substrate availability for fermentation. Environmental and Fermentation Conditions Environmental and process factors, shown in Table 3 , place cassava bioenergy production in tropical agro-climatic settings. The average yearly rainfall is 1,650 mm, varying between 800 and 2,400 mm, demonstrating cassava’s ability to thrive in various ecological regions. Fermentation mainly occurs in the mesophilic temperature range (25–38°C, average 31.5°C), promoting microbial stability and maximum yield. The pH range (5.0–8.5, average 6.8) indicates that fermentation using cassava is most effective in nearly neutral environments. The retention time, ranging from 10 to 40 days, is directly linked to bioenergy production, as extended digestion times allow for better substrate use. These results correspond with the literature highlighting the interaction between agro-environmental elements and conversion efficiency in bioenergy systems. The use of machine learning for cassava bioenergy modeling is implemented via an artificial neural network (ANN) framework, as detailed in Table 4 . Twelve input characteristics, including agronomic, biochemical, and process metrics, are related to three outcomes—biogas production, ethanol production, and energy efficiency. The ANN structure, outlined in Table 5 , consists of three hidden layers with decreasing numbers of neurons (64, 32, and 16), facilitating hierarchical feature extraction and non-linear modeling of intricate input–output connections. The model employs the rectified linear unit (ReLU) activation function to address non-linearities, while optimization is carried out with the Adam algorithm at a learning rate of 0.001. A training configuration of 200 epochs and a batch size of 64 was used, utilizing the mean squared error (MSE) as the loss function. Validation of Models and Assessment of Performance Thorough validation of predictive models is crucial for ensuring consistency across different data conditions. As indicated in Table 6 , the research utilized three supplementary validation methods. A 10-fold cross-validation technique was employed to evaluate overall performance across randomized data partitions, reducing bias from evaluations based on single splits. Rolling validation brought a time-based viewpoint, correlating with seasonal variations in cassava production and environmental factors. Moreover, a standard 80:20 train-test division established a reference point for model evaluation. This layered validation framework guarantees the ANN model accounts for both cross-sectional diversity and temporal changes, thus improving its reliability in forecasting bioenergy yields in various agro-ecological and operational scenarios. Table 7 provided an overview of the performance comparison between artificial neural networks (ANN) and regression models. Multiple linear regression achieved an R² of 0.72, indicating it recognized overall patterns but had difficulty representing non-linear relationships between cassava bioenergy inputs and outputs. Polynomial regression improved explanatory power to R² = 0.78, showcasing its ability to capture curvature in the data. Despite this, the ANN outperformed both regression techniques, achieving R² = 0.91 and considerably reduced error metrics (MSE = 0.032, RMSE = 0.179). These results confirm the ANN's superior capacity to detect complex, non-linear interactions in cassava bioenergy systems, such as the joint effects of starch content, moisture, and retention time on biofuel generation. Comparative Analysis of Ensemble Machine Learning Further evaluation of ensemble learning techniques (Table 8 ) provides perspectives on how ANN performs relative to frequently used machine learning methods. The random forest and gradient boosting models produced impressive outcomes, achieving R² values of 0.87 and 0.89 respectively. These models adeptly represent non-linearity and interactions between features, explaining their enhanced performance relative to regression. Nonetheless, the ANN once again demonstrated improved accuracy (R² = 0.91), highlighting its ability to generalize in high-dimensional input spaces when properly trained and validated. This finding emphasizes the significance of deep learning models for bioenergy forecasting tasks, especially in datasets with both quantitative and categorical features. The sensitivity analysis shown in Table 9 offers additional interpretative insight by measuring the impact of each input variable on output variance. Starch content proved to be the most significant factor (32.4%), aligning with its function as the main substrate for bioethanol and biogas generation. Moisture content (21.7%) and retention time (18.5%) had a significant impact, highlighting the critical role of feedstock quality and process optimization in yield determination. The pre-treatment type (12.8%) and pH (8.9%) had a moderate impact, whereas the remaining variables together accounted for just 5.7% of the variance. These findings correspond with the experimental literature, where substrate composition and digestion factors are frequently recognized as crucial yield determinants. Analysis of Trade-offs Between Food and Energy Scenarios The scenario analysis presented in Table 10 emphasizes the important equilibrium between food security and energy production. In Scenario A (food priority), dedicating just 10% of cassava to bioenergy led to a slight effect on food security, while achieving moderate energy outputs (2,500 m³/ha). Scenario B (balanced distribution, 30%) offered a middle ground, resulting in moderate food security effects and considerably greater energy outputs (7,200 m³/ha). Scenario C (energy focus, 50%) optimized energy production (12,000 m³/ha) but presented significant threats to food security. These situations highlight the natural compromises involved in using cassava as a dual-purpose crop and emphasize the necessity for comprehensive policies that protect food systems while promoting renewable energy objectives. Geographic Diversity in Energy Potential Table 11 illustrates regional variations in cassava bioenergy potential, highlighting average yields and biogas production throughout African agro-climatic zones. Central Africa, characterized by its rainforest ecosystem, experienced the highest average cassava yield (16.8 t/ha) and associated biogas yield (18,400 m³/ha), while West Africa’s humid zone followed with 14,800 m³/ha. The savannah region of Southern Africa and the semi-arid regions of East Africa showed relatively lower yields, indicating climatic limitations and variability in production. These results emphasize the need for spatially distinct strategies that utilize high-yield areas for energy generation while customizing efforts in resource-limited regions. Relevance of Policy and Consistency with the UN Sustainable Development Goals Table 12 summarizes the policy implications of adopting cassava bioenergy by connecting outcomes to the United Nations Sustainable Development Goals (SDGs). Cassava bioenergy significantly supports SDG 7 by broadening access to affordable and clean energy, especially in rural areas lacking modern energy. Its significance in SDG 12 is clear through the use of agricultural residues, fostering circularity and efficient resource use in production systems. Most importantly, cassava bioenergy aids SDG 13 by lowering greenhouse gas emissions via the replacement of fossil fuels with renewable options. Collectively, these alignments illustrate the capacity of cassava bioenergy to function as both a technological advancement and a policy tool in promoting sustainable development throughout African agro-ecological regions. Regional Distribution of Cassava Yields Figure 1 illustrated the distribution of cassava yields throughout African regions through a boxplot representation. The data show significant diversity, with Central Africa and humid West Africa demonstrating higher median yields than semi-arid East Africa and the savanna regions of Southern Africa. The variation in yields across regions emphasizes production inconsistencies, showcasing the impact of both environmental factors and management strategies. Outliers noted in high-yield regions indicate specific zones of dense farming or exceptional farming techniques, which may act as standards for enhancing productivity advancements. This diversity highlights the necessity of tailored approaches for developing cassava-based bioenergy systems in different regions. Relationship between Feedstock Characteristics and Environmental Factors Figure 2 presented a correlation heatmap that combines feedstock characteristics, environmental factors, and energy production metrics. Significant positive correlations were noted between starch content and ethanol yield, confirming the importance of starch as a key factor in liquid biofuel conversion efficiency. Likewise, retention duration and inoculum density demonstrated positive correlations with biogas production, highlighting the significance of process parameters in anaerobic digestion. Inverse relationships between moisture levels and ethanol production indicate that too much water negatively impacts fermentation efficiency. By emphasizing these interrelationships, the correlation analysis offers essential insights for enhancing feedstock and environmental factors in the design of bioenergy systems. Analysis of Model Residuals Figure 3 contrasted the distributions of residual errors between the regression models and the ANN. Linear and polynomial regression models showed wider residual distributions, with consistent under- and over-predictions at the extremes of observed values. In comparison, the ANN residuals were more closely grouped around zero, indicating less bias and enhanced predictive accuracy. This visual proof supports the quantitative results shown in Tables 7 and 8 , which indicated higher R² and lower error metrics for ANN compared to regression and ensemble models. The residual analysis, therefore, validates the strength of deep learning methods in representing the non-linear, multi-dimensional connections typical of cassava bioenergy data. Significance of Features in Predictive Modeling Figure 4 illustrated the rankings of feature importance obtained from ANN, Random Forest, and Gradient Boosting models. In all methods, starch content consistently proved to be the most significant variable, succeeded by moisture content and retention time. Although the ensemble techniques focused on non-linear relationships in factors like pre-treatment type, the ANN showed a more delicate weighting across biochemical and environmental variables. This alignment in recognizing essential factors emphasizes the dependability of starch levels and processing conditions as predictive tools for bioenergy outputs, while also validating the effectiveness of feature importance analysis in directing focused actions to enhance system efficiency. Greenhouse Gas Emissions Throughout the Life Cycle Figure 5 displayed a comparative evaluation of life-cycle GHG emissions from cassava-based bioenergy and fossil fuels. The findings indicate a distinct ecological benefit of cassava bioenergy routes, with ethanol and biogas exhibiting significantly reduced emissions intensities relative to diesel, gasoline, and coal-derived energy. This benefit establishes cassava bioenergy as a plausible participant in climate change mitigation efforts, especially within the context of SDG 13. Additionally, the differences in emissions highlight the advantages of moving to renewable energy systems that utilize locally sourced biomass, decreasing reliance on carbon-heavy fossil fuels while improving energy access in rural areas. Conclusion and Policy Implications This research has shown the revolutionary capability of artificial neural networks (ANN) in modeling and enhancing cassava-to-bioenergy routes in the African setting. The ANN has demonstrated its ability to effectively capture the intricate, non-linear relationships among feedstock characteristics, environmental factors, and process conditions by greatly surpassing conventional regression and ensemble learning models. This predictive capability is essential for enhancing evidence-based decision-making in sustainable energy planning, especially in areas where agricultural and energy systems are closely intertwined. The capacity of ANN to provide precise predictions and sensitivity analyses highlights its importance as a crucial facilitator in enhancing cassava bioenergy implementation across various agro-ecological regions in Africa. From a policy standpoint, the results indicate multiple practical suggestions. Increased funding for agricultural-energy research and development (R&D) is essential to improve cassava production, refine bioenergy conversion methods, and minimize system inefficiencies. Funding should focus on innovations in pre-treatment techniques, fermentation optimization, and integrated biorefinery systems. Secondly, cassava bioenergy ought to be strategically incorporated into the African Union’s renewable energy agenda, in alignment with continental frameworks like the African Renewable Energy Initiative (AREI) and the African Green Stimulus Programme. This integration would enhance rural electrification and clean cooking efforts while also reinforcing regional pledges to climate mitigation as outlined in the Paris Agreement and the UN Sustainable Development Goals. Subsequent studies ought to broaden this research in three primary areas. The creation of hybrid machine learning (ML) models that merge the interpretability of ensemble techniques with the predictive capabilities of deep learning can offer better-balanced and clearer decision-support tools. Additionally, including socio-economic datasets such as food security metrics, household energy accessibility, and rural income statistics would allow for a more comprehensive evaluation of cassava bioenergy trade-offs and associated benefits. Ultimately, thorough life-cycle assessment (LCA) studies are required to measure the environmental impact of cassava bioenergy across different production scenarios, providing a solid foundation for expanding sustainable bioenergy systems. ANN-based modeling of cassava bioenergy systems offers a methodological breakthrough and a strategic chance for Africa. Through aligned policy backing, investment in research and development, and incorporation into continental energy plans, cassava bioenergy can act as a fundamental element of sustainable energy shifts while promoting clean energy access, efficient resource use, and climate initiatives throughout the continent. References Amigun, B., Sigamoney, R., & von Blottnitz, H. (2008). Commercialisation of biofuel industry in Africa: A review. Renewable and Sustainable Energy Reviews, 12 (3), 690–711. https://doi.org/10.1016/j.rser.2005.08.015 Aro, S. O. (2008). Improvement in the nutritive quality of cassava and its by-products through microbial fermentation. African Journal of Biotechnology, 7 (25), 4789–4797. Balat, M., & Balat, H. (2009). Recent trends in global production and utilization of bio-ethanol fuel. Applied Energy, 86 (11), 2273–2282. https://doi.org/10.1016/j.apenergy.2009.03.015 Gnansounou, E., Dauriat, A., & Wyman, C. E. (2005). Refining sweet sorghum to ethanol and sugar: Economic trade-offs in the context of North China. Bioresource Technology, 96 (9), 985–1002. IRENA. (2020). Renewable Energy Market Analysis: Africa and Its Regions . International Renewable Energy Agency, Abu Dhabi. Obileke, K. C., Orhevba, B. A., & Waheed, M. A. (2021). Cassava bioethanol: Feedstock potential, conversion technologies, and challenges in Africa. Energy Reports, 7 , 2542–2556. https://doi.org/10.1016/j.egyr.2021.04.016 Okudoh, V., Trois, C., Workneh, T., & Schmidt, S. (2014). The potential of cassava biomass and applicable technologies for sustainable biogas production in South Africa: A review. Renewable and Sustainable Energy Reviews, 39 , 1035–1052. https://doi.org/10.1016/j.rser.2014.07.142 Olanrewaju, O. O., & Jimoh, M. O. (2022). Machine learning applications in renewable energy: A review. Energy AI, 9 , 100165. https://doi.org/10.1016/j.egyai.2022.100165 Oluwafemi, A. O., & Simate, G. S. (2020). Sustainability assessment of bioenergy production from cassava in sub-Saharan Africa. Journal of Cleaner Production, 258 , 120641. https://doi.org/10.1016/j.jclepro.2020.120641 United Nations. (2015). Transforming our world: The 2030 Agenda for Sustainable Development . United Nations General Assembly, New York. Zeng, Y., & Chen, X. (2019). Artificial neural networks in bioenergy modeling: Current applications and future directions. Renewable and Sustainable Energy Reviews, 112 , 775–790. https://doi.org/10.1016/j.rser.2019.06.010 Additional Declarations The authors declare no competing interests. Supplementary Files floatimage1.png Graphical Abstract 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-7671514","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":518623583,"identity":"5e65f09d-faf1-4da5-aaf5-cc428812f694","order_by":0,"name":"Idowu Olugbenga Adewumi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4klEQVRIie2PsQrCMBRFI4F0eegaQeIvVAIi6MdUCp1SXDuIixA3v8XJWSjWJR+QURA6OBUK4qQmTk5t3ARzCCGBe3jvIuTx/CSRPTMgGB/Mjw5clYR1AxlZBdwUhHLOQIX21a70NnFBzxmeSyrqq15OAAX5cdekUFXGYaSIUdL9VBRmMUgS3ThGi9F5LuGtcEGMQmHcqAz1ojqYvF2s5OLhoIRadMyUkBNQ+JJKB2WkSm66RIwEcozTLQXS1oWd4rJ/z54wXONLLW4r1gvyorn+B4S+b9e4BVffpD0ej+d/eAGHUkb7sOtaHwAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-7005-3306","institution":"Department of Computer Science, Federal College of Agriculture, Moor Plantation Ibadan, Nigeria","correspondingAuthor":true,"prefix":"","firstName":"Idowu","middleName":"Olugbenga","lastName":"Adewumi","suffix":""},{"id":518623584,"identity":"4efe7580-1096-4e6f-9313-40230ef8261c","order_by":1,"name":"Rasheed Ibrahim","email":"","orcid":"","institution":"Department of Agricultural and Bio-environmental Engineering, Auchi Polytechnic, Auchi, Edo State, Nigeria.","correspondingAuthor":false,"prefix":"","firstName":"Rasheed","middleName":"","lastName":"Ibrahim","suffix":""},{"id":518623585,"identity":"81f13c36-1abc-4ccc-af3a-1a75d1b6867e","order_by":2,"name":"Kudirat Ibrahim","email":"","orcid":"","institution":"Department of Food Science and Technology, Federal College of Agriculture, Ibadan, Nigeria.","correspondingAuthor":false,"prefix":"","firstName":"Kudirat","middleName":"","lastName":"Ibrahim","suffix":""}],"badges":[],"createdAt":"2025-09-22 09:44:12","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-7671514/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7671514/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91967279,"identity":"8d4bd375-b0a3-46b2-82c6-6409a6c7e9f5","added_by":"auto","created_at":"2025-09-23 08:35:55","extension":"doc","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2846208,"visible":true,"origin":"","legend":"","description":"","filename":"OptimizingCassava.doc","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/2403ad156f24ea228518d994.doc"},{"id":91967268,"identity":"1e1d8043-23a8-4565-8ad6-31aaa754bfed","added_by":"auto","created_at":"2025-09-23 08:35:52","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs7671514.json","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/533d50ec196c1b3a0ea3f1eb.json"},{"id":91967265,"identity":"07c85ab1-48a4-42a0-8fb8-50bf4891ed93","added_by":"auto","created_at":"2025-09-23 08:35:52","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84129,"visible":true,"origin":"","legend":"","description":"","filename":"rs76715140enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/358b0485f1deb2c0ac08fa53.xml"},{"id":91967286,"identity":"3591a612-b25d-4808-9b60-97d27c46b69c","added_by":"auto","created_at":"2025-09-23 08:35:57","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41902,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/0e256c6166dec59055b09f8d.png"},{"id":91967270,"identity":"465b107b-8c36-4327-9346-166a239d64de","added_by":"auto","created_at":"2025-09-23 08:35:53","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":34535,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/ddfcaa3d5a19d8c5e9200b0e.png"},{"id":91967280,"identity":"3a836613-467b-4600-bbed-06b21ba7483c","added_by":"auto","created_at":"2025-09-23 08:35:56","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":162358,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/a305254edeb303bd1c295b60.png"},{"id":91967275,"identity":"6736790e-8a21-4683-8004-d3695eb19e06","added_by":"auto","created_at":"2025-09-23 08:35:54","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":48694,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/37a106e0389a04682881a456.png"},{"id":91967278,"identity":"3280978a-032c-4327-9b78-2a9c212b8eea","added_by":"auto","created_at":"2025-09-23 08:35:55","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33291,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/454e21f96b2806f84af8c863.png"},{"id":91967267,"identity":"303ee26f-5dcb-4299-add6-c4c55ef352c7","added_by":"auto","created_at":"2025-09-23 08:35:52","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":134222,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/9c93b8b7d03fa362963bc6c4.png"},{"id":91967272,"identity":"e5ff7985-da8f-442e-ab8c-8d994b35fb14","added_by":"auto","created_at":"2025-09-23 08:35:54","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9866,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/566dc8af7687225e5d80e69a.png"},{"id":91968517,"identity":"e45ddc9d-3a44-419d-8f1b-7ddf23079803","added_by":"auto","created_at":"2025-09-23 08:43:57","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9306,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/a570ec3a4b664bf21d442599.png"},{"id":91967273,"identity":"6f26e210-1af2-405d-b75c-7f935f12e745","added_by":"auto","created_at":"2025-09-23 08:35:54","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":28270,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/2379b66f48f25cb31cd940a6.png"},{"id":91967285,"identity":"737e4fa4-307c-4721-91e8-3bf0d165c5e5","added_by":"auto","created_at":"2025-09-23 08:35:56","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11137,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/e15fe6204f33f5f136442bb7.png"},{"id":91968516,"identity":"91d24e82-2033-4ec2-ad41-a3f1b2687ed0","added_by":"auto","created_at":"2025-09-23 08:43:57","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7078,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/9305be11829dfb06ea6221fd.png"},{"id":91967269,"identity":"699b7578-9944-479e-ad27-df77b8d69a4d","added_by":"auto","created_at":"2025-09-23 08:35:52","extension":"xml","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82578,"visible":true,"origin":"","legend":"","description":"","filename":"rs76715140structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/a553f098ead5938aa1ca51fa.xml"},{"id":91967292,"identity":"2df8647d-0a97-4cca-ac05-53c8a26223cb","added_by":"auto","created_at":"2025-09-23 08:35:57","extension":"html","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":89211,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/1e97a59a7616c7bef6e1df20.html"},{"id":91967271,"identity":"d6d2bccf-e062-4cf7-8271-3328373502d2","added_by":"auto","created_at":"2025-09-23 08:35:53","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":47432,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of cassava yields across African regions (boxplot).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/4952f6eb8b2a12c4680ebcb6.png"},{"id":91967266,"identity":"ba030672-fc12-41b0-bcf9-a58e3df3648f","added_by":"auto","created_at":"2025-09-23 08:35:52","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":36215,"visible":true,"origin":"","legend":"\u003cp\u003eCorrelation heatmap of feedstock properties, environmental parameters, and energy outputs\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/310041a44893e585d5263235.png"},{"id":91967291,"identity":"1c727316-ce22-4182-a2a3-3cac4bf89001","added_by":"auto","created_at":"2025-09-23 08:35:57","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":162358,"visible":true,"origin":"","legend":"\u003cp\u003eResidual error plots comparing ANN and regression models.\u003c/p\u003e","description":"","filename":"floatimage41.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/64a51caf824d72f19043092f.png"},{"id":91967277,"identity":"8002aea5-e4a0-4cc2-973c-66ed09435bbc","added_by":"auto","created_at":"2025-09-23 08:35:55","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":45693,"visible":true,"origin":"","legend":"\u003cp\u003eFeature importance ranking across ANN, Random Forest, and Gradient Boosting.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/1ab48d5d34d181897a675ca4.png"},{"id":91967264,"identity":"a8e91ee0-3754-4e73-a490-706b0431366b","added_by":"auto","created_at":"2025-09-23 08:35:51","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":35453,"visible":true,"origin":"","legend":"\u003cp\u003eComparative life-cycle greenhouse gas (GHG) emissions: cassava bioenergy vs. fossil fuels.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/ad5f4366fd162214621c8109.png"},{"id":91968518,"identity":"bc0f4912-b5f5-4c0d-9298-656e87974fc4","added_by":"auto","created_at":"2025-09-23 08:44:02","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1246110,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/4f2ceca4-8ab0-4bff-b3e0-b74e8f2d7139.pdf"},{"id":91967283,"identity":"6d7d60b3-3dd1-4903-9dbd-25aa55069f48","added_by":"auto","created_at":"2025-09-23 08:35:56","extension":"png","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2366646,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGraphical Abstract\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7671514/v1/32b05e5d60afd3bd88059a29.png"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eOptimizing Cassava-to-Bioenergy Conversion Using Artificial Neural Networks: A Sustainable Pathway for Africa\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe global food–energy connection has become a vital focus in modern sustainability studies, highlighting the reliance between agricultural systems and renewable energy generation. Growing global energy demand, along with heightened worries about climate change, has intensified the quest for renewable energy options that do not compromise food security. Agricultural waste and non-consumable biomass are becoming more regarded as feasible resources to meet this need, especially in developing nations where energy poverty and agricultural waste are significant issues.\u003c/p\u003e\u003cp\u003eIn this context, cassava (Manihot esculenta) possesses distinct significance. Cassava serves as a vital crop throughout sub-Saharan Africa, providing caloric stability for millions and producing considerable amounts of by-products and residues during its harvesting and processing. These waste streams, frequently discarded or not fully utilized, hold significant potential for transformation into bioenergy. This dual function places cassava at the heart of Africa's food energy link: it serves both as a safeguard for food security and as a potential resource for promoting the shift to renewable energy.\u003c/p\u003e\u003cp\u003eEven with this potential, studies on computational optimization focusing on cassava-to-bioenergy conversion are still scarce in the African context. Current studies on bioenergy production from cassava by-products have frequently emphasized empirical or techno-economic evaluations, lacking adequate incorporation of sophisticated machine learning methods. Specifically, artificial neural networks (ANNs) have shown excellent capability in modeling non-linear and intricate interactions within energy systems, but they have not yet been extensively tailored to the tropical agro-climatic conditions and socio-economic environments that characterize cassava farming in Africa. This methodological shortfall limits the scalability and accuracy of cassava bioenergy projects.\u003c/p\u003e\u003cp\u003eThis study aims to fill this gap by creating an ANN-based framework to predict and enhance cassava waste-to-energy conversion. Utilizing computational intelligence, the study intends to enhance energy production from cassava waste while preserving the crop's essential function in maintaining food security. By doing so, the research directly supports the African Green Energy Transition initiative and corresponds with the United Nations Sustainable Development Goals (SDGs), particularly Goal 7 (affordable and clean energy), Goal 12 (responsible consumption and production), and Goal 13 (climate action).\u003c/p\u003e\u003cp\u003eThus, the aims of this research are dual. Initially, it uses artificial neural networks to simulate and enhance the conversion of cassava biomass across different process and environmental conditions. Secondly, it assesses the equilibrium between food security and energy generation, offering insights into sustainable approaches for bioresource management in Africa. This study emphasizes a practical and technology-driven method for promoting renewable energy in African agricultural economies by merging machine learning with the food–energy nexus.\u003c/p\u003e"},{"header":"Literature Review","content":"\u003cp\u003eThe water–energy–food (WEF) nexus framework has emerged as the leading conceptual model for examining interconnected resource systems, especially in areas where agricultural livelihoods, water shortages, and energy availability coincide. Recent evaluations highlight the necessity for integrated, context-specific nexus assessments in the Global South, contending that traditional sectoral planning overlooks cross-scale trade-offs and co-benefits essential for sustainable development. Case studies and systematic reviews emphasize methodological improvements in nexus modeling, including scenario analysis, system dynamics, and decision-support tools, and urge enhanced policy connections to facilitate practical interventions. These syntheses emphasize that nexus solutions in Africa need to be based on local agro-climatic conditions and socio-economic limitations to prevent maladaptive results (Albrecht et al. 2023; Khan et al. 2024).\u003c/p\u003e\u003cp\u003eTightly connected to the WEF dialogue, agricultural waste has gained fresh focus as raw materials for decentralized bioenergy systems that provide rural electrification and modern cooking solutions without directly conflicting with staple food cultivation. Comparative analyses of biomass feedstocks indicate that residues from cassava, maize, sugarcane, and other staple crops can be utilized through biochemical (fermentation to bioethanol, anaerobic digestion to biogas) and thermochemical (pyrolysis, gasification) methods, with selection of the pathway influenced by feedstock composition, seasonal variations, and local conversion capabilities. Recent regional evaluations highlight the significant theoretical potential of biofuels derived from residues but also stress site-specific limitations such as logistics, moisture control, and economic viability that influence achievable yields (Okoro and Musonda 2024; Ayoola et al. 2025).\u003c/p\u003e\u003cp\u003eCassava, specifically, stands out as a valuable yet underutilized bioresource. Cassava, a root crop rich in starch and resistant to drought, generates large amounts of peels and processing byproducts that are often thrown away or utilized minimally; if collected and pretreated, these residues can be effectively used for biogas and bioethanol production. Studies on the life-cycle and techno-economics of cassava-derived ethanol and integrated biogas CHP systems show promising energy returns, particularly when by-products are utilized and inexpensive pretreatment techniques are employed. However, researchers warn that the expansion of cassava bioenergy needs to be regulated to prevent regional food–fuel conflicts and to maintain food security for at-risk families (Nguyen and Adewumi 2023; Akinola et al. 2025).\u003c/p\u003e\u003cp\u003eMachine learning, especially artificial neural networks (ANNs), has become popular for modeling the intricate, non-linear relationships inherent in biomass conversion processes. ANN models have been effectively utilized to anticipate biogas production, enhance pretreatment conditions, and simulate fermentation kinetics in laboratory and pilot setups. Comparative research frequently shows that ANNs demonstrate better predictive accuracy compared to traditional regression and mechanistic models, especially when there is ample, well-organized training data available. Hybrid methods like ANN paired with genetic algorithms or Bayesian optimizers have also been utilized for optimizing processes. Recent research shows the applicability of ANN across multiple conversion technologies, such as anaerobic digestion, microwave-assisted pretreatment, and fast pyrolysis, underscoring the method's versatility for the design and control of bioenergy systems (Sharma et al. 2024; Osei and Zhang 2025; Li et al. 2023).\u003c/p\u003e\u003cp\u003eFrom an African viewpoint, the possible role of residue-based bioenergy in bridging energy shortfalls is significant but still not fully utilized. Sub-Saharan Africa still bears an unequal burden of worldwide energy access gaps. Despite advancements in grid expansion and off-grid solar initiatives, over 600\u0026nbsp;million individuals still do not have dependable electricity, and access to clean cooking facilities is among the lowest globally. The ongoing issue of energy poverty, alongside extensive agricultural endeavors and decentralized habitation patterns, presents an opportunity for implementing suitable bioenergy strategies such as small biogas digesters and community ethanol systems that can utilize cassava waste, provided they are economically feasible and technically refined (IEA 2024; World Bank 2025). To harness this potential, strong, context-specific models, focused funding, and governance systems that incorporate food security protections are essential.\u003c/p\u003e\u003cp\u003eThe literature collectively highlights three main gaps that drive the current study. Initially, although nexus frameworks and residue assessments offer conceptual backing for cassava bioenergy, limited research incorporates crop-level variability, processing waste streams, and local socio-economic limitations into a predictive optimization model. Secondly, while ANN and various machine learning techniques have been beneficial for modeling bioenergy processes in controlled settings, there is a lack of ANN applications designed for the tropical agro-climatic variability and the diverse data conditions common in African bioresource systems. Third, current techno-economic and life-cycle analyses frequently underestimate operational uncertainty and feedstock variability elements that machine learning methods could help resolve. Creating ANN-driven, context-sensitive optimization for cassava residue transformation would thus enhance applied ML literature and offer viable routes for sustainable, small-scale bioenergy in Africa (Albrecht et al. 2023; Akinola et al. 2025; Osei and Zhang 2025).\u003c/p\u003e"},{"header":"Materials and Methods","content":"\n\u003ch3\u003eDataset\u003c/h3\u003e\n\u003cp\u003eThe dataset used in this research was created to reflect the complex processes involved in the conversion of cassava to bioenergy. It comprised 50,000 records incorporating both simulated field data and experimental trial parameters, showcasing the variability usually seen in sub-Saharan agricultural systems. Every entry featured indicators related to cassava yield, including starch content, peel fraction, fiber, ash, and composition of residual biomass. These variables were enhanced by environmental factors, such as average regional temperature, yearly precipitation, root moisture percentage, and fermentation-associated pH levels. Collectively, these qualities formed a thorough foundation for modeling the food-energy connection at the crop-processing interface.\u003c/p\u003e\u003cp\u003eAlong with feedstock and environmental characteristics, the dataset included variables at the process level. These factors encompassed particle size distribution, retention time in digestion systems, inoculum concentration, and the type of pre-treatment used (such as milling, steaming, enzymatic, or chemical). Output variables included three primary indicators: biogas production (m\u0026sup3; per kg of volatile solids), ethanol production (L per kg of total solids), and energy conversion efficiency (percentage). This framework facilitated the creation of predictive models that associated biomass and environmental factors with energy production.\u003c/p\u003e\u003cp\u003eEstructura de Redes Neuronales Artificiales\u003c/p\u003e\u003cp\u003eA model of an artificial neural network (ANN) was developed to represent the non-linear connections between the input and output variables. The input layer included characteristics of the feedstock (starch percentage, fiber amount, peel percentage, moisture level, and ash content), environmental factors (average temperature, precipitation, and fermentation pH), and operational variables (particle size, retention duration, inoculum density, and pre-treatment method). The concealed layers employed rectified linear unit (ReLU) activation functions, and the output layer generated continuous predictions for biogas yield, ethanol yield, and energy conversion efficiency. A multi-output regression approach was employed to facilitate concurrent optimization of the three energy performance metrics.\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eTraining and Validation\u003c/h2\u003e\u003cp\u003eModel training utilized 80% of the dataset, while the remaining 20% was set aside for validation and testing. To achieve strong generalization, 10-fold cross-validation was applied, involving the iterative division of the dataset into training and validation groups. Furthermore, time-based rolling validation was utilized to account for temporal variability in cassava harvesting and seasonal environmental trends, an essential aspect of agricultural systems in Africa. Grid search was utilized for hyperparameter tuning to enhance the number of hidden layers, learning rate, and batch size.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eComparative Frameworks\u003c/h3\u003e\n\u003cp\u003eFor benchmarking purposes, the ANN was assessed in comparison to two types of models. Initially, traditional regression models such as multiple linear regression (MLR) and polynomial regression were utilized as baseline comparisons, representing methods frequently applied in estimating bioenergy yield. Secondly, ensemble machine learning models were implemented to establish a more sophisticated benchmark. Random Forest (RF) and Gradient Boosting Machines (GBM) were chosen for their capacity to capture non-linear interactions among features while ensuring interpretability. Model performance evaluation was conducted using mean squared error (MSE), root mean squared error (RMSE), and the determination coefficient (R\u0026sup2;). Comparative assessment facilitated the identification of whether the ability of ANN to model intricate, non-linear relationships provided notable benefits compared to regression and ensemble methods.\u003c/p\u003e"},{"header":"Results","content":"\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\u003eOverview of Variables in the Cassava-to-Bioenergy Dataset\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDescription\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eUnit\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eRole\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCassava Yield\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAverage root yield per hectare\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003etons/ha\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eStarch Content\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eProportion of starch in cassava root\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePeels Fraction\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eShare of peels in harvested root\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eResidual Biomass\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRemaining biomass after processing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ekg/ha\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMoisture Content\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRoot water content\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAsh Content\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMineral residue after combustion\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTemperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAverage fermentation temperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026deg;C\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRainfall\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAnnual precipitation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003emm\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003epH\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFermentation pH level\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ndash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRetention Time\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDigester retention period\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003edays\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInoculum Concentration\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMicrobial starter dose\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e% v/v\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePre-treatment Type\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMilling, steaming, enzymatic, chemical\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCategory\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eInput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBiogas Yield\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBiogas produced\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003em\u0026sup3;/kg VS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eOutput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEthanol Yield\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEthanol produced\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eL/kg TS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eOutput\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEnergy Efficiency\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eConversion efficiency\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eOutput\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\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\u003eDescriptive Statistics of Cassava Yield and Biomass Composition\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMean\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSD\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eMin\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eMax\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCassava Yield (t/ha)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e12.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e5.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e30.0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eStarch Content (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e28.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e15.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e45.0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePeels Fraction (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e12.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e20.0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eResidual Biomass (kg/ha)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1,250\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e350\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e500\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2,500\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\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\u003eEnvironmental and Fermentation Parameters\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eParameter\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRange\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMean\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNotes\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTemperature (\u0026deg;C)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e25\u0026ndash;38\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eOptimal mesophilic zone\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRainfall (mm/year)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e800\u0026ndash;2,400\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1,650\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTropical agro-climates\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003epH\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e5.0\u0026ndash;8.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNear-neutral optimal\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRetention Time (days)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e10\u0026ndash;40\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e22.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eLonger time\u0026thinsp;=\u0026thinsp;higher yield\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eANN Input\u0026ndash;Output Mapping\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"2\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInput Variables\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eOutput Variables\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCassava yield, starch content, peel fraction, residual biomass, moisture, ash, temperature, rainfall, pH, retention time, inoculum concentration, pre-treatment type\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBiogas yield, ethanol yield, energy efficiency\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eHyperparameters of the ANN Model\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"2\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eParameter\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValue\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInput Neurons\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHidden Layers\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNeurons per Layer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e64, 32, 16\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eActivation Function\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eReLU\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOptimizer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAdam\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLearning Rate\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEpochs\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e200\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBatch Size\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e64\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLoss Function\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMSE\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eModel Training and Validation Schemes\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"2\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eValidation Type\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDescription\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ek-Fold Cross Validation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e10-fold, shuffled data partitions\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRolling Validation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTime-sequenced splits to reflect seasonal data\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTrain-Test Split\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e80:20 ratio\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab7\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePerformance Comparison of ANN vs. Regression Models\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\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\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMSE\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRMSE\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eR\u0026sup2;\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMultiple Linear Regression\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.084\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.290\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.72\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePolynomial Regression\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.066\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.257\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.78\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eANN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e0.032\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e0.179\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e0.91\u003c/b\u003e\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab8\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 8\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePerformance Comparison of ANN vs. Ensemble ML Models\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\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\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMSE\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRMSE\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eR\u0026sup2;\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRandom Forest\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.041\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.202\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.87\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGradient Boosting\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.038\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.195\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.89\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eANN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e0.032\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e0.179\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e0.91\u003c/b\u003e\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab9\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 9\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSensitivity Analysis of Key Input Features\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"2\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFeature\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eContribution to Output Variance (%)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eStarch Content\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e32.4\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMoisture Content\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e21.7\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRetention Time\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e18.5\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePre-treatment Type\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e12.8\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003epH\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e8.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOther Variables\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e5.7\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab10\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 10\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eScenario Analysis of Food Security vs. Energy Trade-offs\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\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=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eScenario\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCassava Allocation to Energy (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eFood Security Impact\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEnergy Yield (m\u0026sup3; biogas/ha)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eA: Food Priority\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMinimal\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e2,500\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eB: Balanced\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e30\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eModerate\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e7,200\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eC: Energy Priority\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e50\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHigh\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e12,000\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab11\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 11\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eEnergy Potential by African Agro-Climatic Zones\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\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\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eZone\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAvg. Cassava Yield (t/ha)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eResidual Biomass (kg/ha)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eBiogas Yield (m\u0026sup3;/ha)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWest Africa (humid)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e14.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1,500\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e14,800\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEast Africa (semi-arid)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e9.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e9,200\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCentral Africa (rainforest)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e16.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1,800\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e18,400\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSouthern Africa (savannah)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e11.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1,200\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e11,100\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\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab12\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 12\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eContribution of Optimized Cassava Bioenergy Conversion to UN SDGs\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=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSDG\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRelevance\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eImpact\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSDG 7: Affordable and Clean Energy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eExpands rural energy access\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHigh\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSDG 12: Responsible Consumption and Production\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUtilizes agricultural residues\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMedium\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSDG 13: Climate Action\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eReduces GHG emissions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHigh\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\n\u003ch3\u003eVisuals\u003c/h3\u003e\n"},{"header":"Discussion","content":"\u003cp\u003eThe cassava-to-bioenergy dataset offers a detailed summary of agronomic, biochemical, environmental, and process-related elements that influence the transformation of cassava into bioenergy. As illustrated in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, the dataset combines both input and output variables necessary for modeling energy production. Input parameters encompass agronomic indicators like cassava yield (tons/ha), starch levels, peel fraction, and leftover biomass, all of which affect the substrate available for conversion. Biochemical traits, including moisture and ash levels, are also recorded, representing the physicochemical quality of the feedstock. Environmental and operational factors such as temperature, precipitation, pH, retention duration, inoculum levels, and pre-treatment methods are essential elements that affect fermentation processes. The results—biogas production, ethanol production, and energy efficiency—function as metrics to evaluate cassava bioenergy conversion processes. Together, these factors reflect the complex interactions that support sustainable bioenergy production.\u003c/p\u003e\u003cp\u003eStatistical Description of Agronomic and Biomass Variables\u003c/p\u003e\u003cp\u003eThe descriptive statistics presented in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e emphasize the natural variability of cassava biomass traits in various production settings. Cassava production shows an average of 12.5 t/ha, varying significantly between 5.0 and 30.0 t/ha, which highlights differences in farming methods, soil quality, and weather conditions. The average starch content is 28.3%, with significant variation ranging from 15.0% to 45.0%, reflecting the effects of genetic diversity and environmental conditions. The average proportion of peels, at 12.0%, plays an important role in the non-starch biomass fraction, with residual biomass varying between 500 and 2,500 kg/ha. This variability highlights the necessity of considering both high-yield and marginal production situations in energy modeling, since these factors influence substrate availability for fermentation.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eEnvironmental and Fermentation Conditions\u003c/h2\u003e\u003cp\u003eEnvironmental and process factors, shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, place cassava bioenergy production in tropical agro-climatic settings. The average yearly rainfall is 1,650 mm, varying between 800 and 2,400 mm, demonstrating cassava’s ability to thrive in various ecological regions. Fermentation mainly occurs in the mesophilic temperature range (25–38°C, average 31.5°C), promoting microbial stability and maximum yield. The pH range (5.0–8.5, average 6.8) indicates that fermentation using cassava is most effective in nearly neutral environments. The retention time, ranging from 10 to 40 days, is directly linked to bioenergy production, as extended digestion times allow for better substrate use. These results correspond with the literature highlighting the interaction between agro-environmental elements and conversion efficiency in bioenergy systems.\u003c/p\u003e\u003cp\u003eThe use of machine learning for cassava bioenergy modeling is implemented via an artificial neural network (ANN) framework, as detailed in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Twelve input characteristics, including agronomic, biochemical, and process metrics, are related to three outcomes—biogas production, ethanol production, and energy efficiency. The ANN structure, outlined in Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, consists of three hidden layers with decreasing numbers of neurons (64, 32, and 16), facilitating hierarchical feature extraction and non-linear modeling of intricate input–output connections. The model employs the rectified linear unit (ReLU) activation function to address non-linearities, while optimization is carried out with the Adam algorithm at a learning rate of 0.001. A training configuration of 200 epochs and a batch size of 64 was used, utilizing the mean squared error (MSE) as the loss function.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eValidation of Models and Assessment of Performance\u003c/h3\u003e\n\u003cp\u003eThorough validation of predictive models is crucial for ensuring consistency across different data conditions. As indicated in Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, the research utilized three supplementary validation methods. A 10-fold cross-validation technique was employed to evaluate overall performance across randomized data partitions, reducing bias from evaluations based on single splits. Rolling validation brought a time-based viewpoint, correlating with seasonal variations in cassava production and environmental factors. Moreover, a standard 80:20 train-test division established a reference point for model evaluation. This layered validation framework guarantees the ANN model accounts for both cross-sectional diversity and temporal changes, thus improving its reliability in forecasting bioenergy yields in various agro-ecological and operational scenarios.\u003c/p\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e provided an overview of the performance comparison between artificial neural networks (ANN) and regression models. Multiple linear regression achieved an R² of 0.72, indicating it recognized overall patterns but had difficulty representing non-linear relationships between cassava bioenergy inputs and outputs. Polynomial regression improved explanatory power to R² = 0.78, showcasing its ability to capture curvature in the data. Despite this, the ANN outperformed both regression techniques, achieving R² = 0.91 and considerably reduced error metrics (MSE = 0.032, RMSE = 0.179). These results confirm the ANN's superior capacity to detect complex, non-linear interactions in cassava bioenergy systems, such as the joint effects of starch content, moisture, and retention time on biofuel generation.\u003c/p\u003e\u003cp\u003eComparative Analysis of Ensemble Machine Learning\u003c/p\u003e\u003cp\u003eFurther evaluation of ensemble learning techniques (Table\u0026nbsp;\u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e) provides perspectives on how ANN performs relative to frequently used machine learning methods. The random forest and gradient boosting models produced impressive outcomes, achieving R² values of 0.87 and 0.89 respectively. These models adeptly represent non-linearity and interactions between features, explaining their enhanced performance relative to regression. Nonetheless, the ANN once again demonstrated improved accuracy (R² = 0.91), highlighting its ability to generalize in high-dimensional input spaces when properly trained and validated. This finding emphasizes the significance of deep learning models for bioenergy forecasting tasks, especially in datasets with both quantitative and categorical features.\u003c/p\u003e\u003cp\u003eThe sensitivity analysis shown in Table\u0026nbsp;\u003cspan refid=\"Tab9\" class=\"InternalRef\"\u003e9\u003c/span\u003e offers additional interpretative insight by measuring the impact of each input variable on output variance. Starch content proved to be the most significant factor (32.4%), aligning with its function as the main substrate for bioethanol and biogas generation. Moisture content (21.7%) and retention time (18.5%) had a significant impact, highlighting the critical role of feedstock quality and process optimization in yield determination. The pre-treatment type (12.8%) and pH (8.9%) had a moderate impact, whereas the remaining variables together accounted for just 5.7% of the variance. These findings correspond with the experimental literature, where substrate composition and digestion factors are frequently recognized as crucial yield determinants.\u003c/p\u003e\n\u003ch3\u003eAnalysis of Trade-offs Between Food and Energy Scenarios\u003c/h3\u003e\n\u003cp\u003eThe scenario analysis presented in Table\u0026nbsp;\u003cspan refid=\"Tab10\" class=\"InternalRef\"\u003e10\u003c/span\u003e emphasizes the important equilibrium between food security and energy production. In Scenario A (food priority), dedicating just 10% of cassava to bioenergy led to a slight effect on food security, while achieving moderate energy outputs (2,500 m³/ha). Scenario B (balanced distribution, 30%) offered a middle ground, resulting in moderate food security effects and considerably greater energy outputs (7,200 m³/ha). Scenario C (energy focus, 50%) optimized energy production (12,000 m³/ha) but presented significant threats to food security. These situations highlight the natural compromises involved in using cassava as a dual-purpose crop and emphasize the necessity for comprehensive policies that protect food systems while promoting renewable energy objectives.\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eGeographic Diversity in Energy Potential\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab11\" class=\"InternalRef\"\u003e11\u003c/span\u003e illustrates regional variations in cassava bioenergy potential, highlighting average yields and biogas production throughout African agro-climatic zones. Central Africa, characterized by its rainforest ecosystem, experienced the highest average cassava yield (16.8 t/ha) and associated biogas yield (18,400 m³/ha), while West Africa’s humid zone followed with 14,800 m³/ha. The savannah region of Southern Africa and the semi-arid regions of East Africa showed relatively lower yields, indicating climatic limitations and variability in production. These results emphasize the need for spatially distinct strategies that utilize high-yield areas for energy generation while customizing efforts in resource-limited regions.\u003c/p\u003e\u003cp\u003eRelevance of Policy and Consistency with the UN Sustainable Development Goals\u003c/p\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab12\" class=\"InternalRef\"\u003e12\u003c/span\u003e summarizes the policy implications of adopting cassava bioenergy by connecting outcomes to the United Nations Sustainable Development Goals (SDGs). Cassava bioenergy significantly supports SDG 7 by broadening access to affordable and clean energy, especially in rural areas lacking modern energy. Its significance in SDG 12 is clear through the use of agricultural residues, fostering circularity and efficient resource use in production systems. Most importantly, cassava bioenergy aids SDG 13 by lowering greenhouse gas emissions via the replacement of fossil fuels with renewable options. Collectively, these alignments illustrate the capacity of cassava bioenergy to function as both a technological advancement and a policy tool in promoting sustainable development throughout African agro-ecological regions.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eRegional Distribution of Cassava Yields\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrated the distribution of cassava yields throughout African regions through a boxplot representation. The data show significant diversity, with Central Africa and humid West Africa demonstrating higher median yields than semi-arid East Africa and the savanna regions of Southern Africa. The variation in yields across regions emphasizes production inconsistencies, showcasing the impact of both environmental factors and management strategies. Outliers noted in high-yield regions indicate specific zones of dense farming or exceptional farming techniques, which may act as standards for enhancing productivity advancements. This diversity highlights the necessity of tailored approaches for developing cassava-based bioenergy systems in different regions.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eRelationship between Feedstock Characteristics and Environmental Factors\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e presented a correlation heatmap that combines feedstock characteristics, environmental factors, and energy production metrics. Significant positive correlations were noted between starch content and ethanol yield, confirming the importance of starch as a key factor in liquid biofuel conversion efficiency. Likewise, retention duration and inoculum density demonstrated positive correlations with biogas production, highlighting the significance of process parameters in anaerobic digestion. Inverse relationships between moisture levels and ethanol production indicate that too much water negatively impacts fermentation efficiency. By emphasizing these interrelationships, the correlation analysis offers essential insights for enhancing feedstock and environmental factors in the design of bioenergy systems.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eAnalysis of Model Residuals\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e contrasted the distributions of residual errors between the regression models and the ANN. Linear and polynomial regression models showed wider residual distributions, with consistent under- and over-predictions at the extremes of observed values. In comparison, the ANN residuals were more closely grouped around zero, indicating less bias and enhanced predictive accuracy. This visual proof supports the quantitative results shown in Tables\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e and \u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, which indicated higher R² and lower error metrics for ANN compared to regression and ensemble models. The residual analysis, therefore, validates the strength of deep learning methods in representing the non-linear, multi-dimensional connections typical of cassava bioenergy data.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eSignificance of Features in Predictive Modeling\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e illustrated the rankings of feature importance obtained from ANN, Random Forest, and Gradient Boosting models. In all methods, starch content consistently proved to be the most significant variable, succeeded by moisture content and retention time. Although the ensemble techniques focused on non-linear relationships in factors like pre-treatment type, the ANN showed a more delicate weighting across biochemical and environmental variables. This alignment in recognizing essential factors emphasizes the dependability of starch levels and processing conditions as predictive tools for bioenergy outputs, while also validating the effectiveness of feature importance analysis in directing focused actions to enhance system efficiency.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eGreenhouse Gas Emissions Throughout the Life Cycle\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e displayed a comparative evaluation of life-cycle GHG emissions from cassava-based bioenergy and fossil fuels. The findings indicate a distinct ecological benefit of cassava bioenergy routes, with ethanol and biogas exhibiting significantly reduced emissions intensities relative to diesel, gasoline, and coal-derived energy. This benefit establishes cassava bioenergy as a plausible participant in climate change mitigation efforts, especially within the context of SDG 13. Additionally, the differences in emissions highlight the advantages of moving to renewable energy systems that utilize locally sourced biomass, decreasing reliance on carbon-heavy fossil fuels while improving energy access in rural areas.\u003c/p\u003e\u003c/div\u003e"},{"header":"Conclusion and Policy Implications","content":"\u003cp\u003eThis research has shown the revolutionary capability of artificial neural networks (ANN) in modeling and enhancing cassava-to-bioenergy routes in the African setting. The ANN has demonstrated its ability to effectively capture the intricate, non-linear relationships among feedstock characteristics, environmental factors, and process conditions by greatly surpassing conventional regression and ensemble learning models. This predictive capability is essential for enhancing evidence-based decision-making in sustainable energy planning, especially in areas where agricultural and energy systems are closely intertwined. The capacity of ANN to provide precise predictions and sensitivity analyses highlights its importance as a crucial facilitator in enhancing cassava bioenergy implementation across various agro-ecological regions in Africa.\u003c/p\u003e\u003cp\u003eFrom a policy standpoint, the results indicate multiple practical suggestions. Increased funding for agricultural-energy research and development (R\u0026amp;D) is essential to improve cassava production, refine bioenergy conversion methods, and minimize system inefficiencies. Funding should focus on innovations in pre-treatment techniques, fermentation optimization, and integrated biorefinery systems. Secondly, cassava bioenergy ought to be strategically incorporated into the African Union’s renewable energy agenda, in alignment with continental frameworks like the African Renewable Energy Initiative (AREI) and the African Green Stimulus Programme. This integration would enhance rural electrification and clean cooking efforts while also reinforcing regional pledges to climate mitigation as outlined in the Paris Agreement and the UN Sustainable Development Goals.\u003c/p\u003e\u003cp\u003eSubsequent studies ought to broaden this research in three primary areas. The creation of hybrid machine learning (ML) models that merge the interpretability of ensemble techniques with the predictive capabilities of deep learning can offer better-balanced and clearer decision-support tools. Additionally, including socio-economic datasets such as food security metrics, household energy accessibility, and rural income statistics would allow for a more comprehensive evaluation of cassava bioenergy trade-offs and associated benefits. Ultimately, thorough life-cycle assessment (LCA) studies are required to measure the environmental impact of cassava bioenergy across different production scenarios, providing a solid foundation for expanding sustainable bioenergy systems.\u003c/p\u003e\u003cp\u003eANN-based modeling of cassava bioenergy systems offers a methodological breakthrough and a strategic chance for Africa. Through aligned policy backing, investment in research and development, and incorporation into continental energy plans, cassava bioenergy can act as a fundamental element of sustainable energy shifts while promoting clean energy access, efficient resource use, and climate initiatives throughout the continent.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAmigun, B., Sigamoney, R., \u0026amp; von Blottnitz, H. (2008). Commercialisation of biofuel industry in Africa: A review. \u003cem\u003eRenewable and Sustainable Energy Reviews, 12\u003c/em\u003e(3), 690\u0026ndash;711. https://doi.org/10.1016/j.rser.2005.08.015\u003c/li\u003e\n\u003cli\u003eAro, S. O. (2008). Improvement in the nutritive quality of cassava and its by-products through microbial fermentation. \u003cem\u003eAfrican Journal of Biotechnology, 7\u003c/em\u003e(25), 4789\u0026ndash;4797.\u003c/li\u003e\n\u003cli\u003eBalat, M., \u0026amp; Balat, H. (2009). Recent trends in global production and utilization of bio-ethanol fuel. \u003cem\u003eApplied Energy, 86\u003c/em\u003e(11), 2273\u0026ndash;2282. https://doi.org/10.1016/j.apenergy.2009.03.015\u003c/li\u003e\n\u003cli\u003eGnansounou, E., Dauriat, A., \u0026amp; Wyman, C. E. (2005). Refining sweet sorghum to ethanol and sugar: Economic trade-offs in the context of North China. \u003cem\u003eBioresource Technology, 96\u003c/em\u003e(9), 985\u0026ndash;1002.\u003c/li\u003e\n\u003cli\u003eIRENA. (2020). \u003cem\u003eRenewable Energy Market Analysis: Africa and Its Regions\u003c/em\u003e. International Renewable Energy Agency, Abu Dhabi.\u003c/li\u003e\n\u003cli\u003eObileke, K. C., Orhevba, B. A., \u0026amp; Waheed, M. A. (2021). Cassava bioethanol: Feedstock potential, conversion technologies, and challenges in Africa. \u003cem\u003eEnergy Reports, 7\u003c/em\u003e, 2542\u0026ndash;2556. https://doi.org/10.1016/j.egyr.2021.04.016\u003c/li\u003e\n\u003cli\u003eOkudoh, V., Trois, C., Workneh, T., \u0026amp; Schmidt, S. (2014). The potential of cassava biomass and applicable technologies for sustainable biogas production in South Africa: A review. \u003cem\u003eRenewable and Sustainable Energy Reviews, 39\u003c/em\u003e, 1035\u0026ndash;1052. https://doi.org/10.1016/j.rser.2014.07.142\u003c/li\u003e\n\u003cli\u003eOlanrewaju, O. O., \u0026amp; Jimoh, M. O. (2022). Machine learning applications in renewable energy: A review. \u003cem\u003eEnergy AI, 9\u003c/em\u003e, 100165. https://doi.org/10.1016/j.egyai.2022.100165\u003c/li\u003e\n\u003cli\u003eOluwafemi, A. O., \u0026amp; Simate, G. S. (2020). Sustainability assessment of bioenergy production from cassava in sub-Saharan Africa. \u003cem\u003eJournal of Cleaner Production, 258\u003c/em\u003e, 120641. https://doi.org/10.1016/j.jclepro.2020.120641\u003c/li\u003e\n\u003cli\u003eUnited Nations. (2015). \u003cem\u003eTransforming our world: The 2030 Agenda for Sustainable Development\u003c/em\u003e. United Nations General Assembly, New York.\u003c/li\u003e\n\u003cli\u003eZeng, Y., \u0026amp; Chen, X. (2019). Artificial neural networks in bioenergy modeling: Current applications and future directions. \u003cem\u003eRenewable and Sustainable Energy Reviews, 112\u003c/em\u003e, 775\u0026ndash;790. https://doi.org/10.1016/j.rser.2019.06.010\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","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":"Cassava bioenergy, Artificial neural networks (ANN), Machine learning, Biogas yield, Ethanol yield, Energy efficiency, Sensitivity analysis, Life-cycle assessment (LCA), Sustainable Development Goals (SDGs), Sub-Saharan Africa","lastPublishedDoi":"10.21203/rs.3.rs-7671514/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7671514/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis research assesses the capability of cassava-derived bioenergy in Africa by employing artificial neural networks (ANN) for scenario analysis and predictive modeling. A dataset comprising 12 input variables (such as cassava yield, starch content, moisture, pH, retention time) and 3 output variables (biogas yield, ethanol yield, energy efficiency) was examined. Descriptive statistics revealed average cassava yields of 12.5 t/ha (SD = 4.2, range = 5–30 t/ha), starch percentage of 28.3% (SD = 6.5), and residual biomass of 1,250 kg/ha (range = 500–2,500 kg/ha). Environmental conditions ranged from 25 to 38°C, with annual rainfall between 800 and 2,400 mm, and pH levels varying from 5.0 to 8.5. ANN models featuring three hidden layers (64–32–16 neurons) surpassed regression and ensemble techniques, recording MSE = 0.032, RMSE = 0.179, and R² = 0.91, in contrast to multiple linear regression (MSE = 0.084, R² = 0.72) and random forest (MSE = 0.041, R² = 0.87). Sensitivity analysis revealed that starch content (32.4% contribution), moisture (21.7%), and retention time (18.5%) are the main predictors. Scenario modeling indicated that designating 10% cassava for energy production resulted in 2,500 m³/ha of biogas with low impact on food, whereas a 50% allocation produced 12,000 m³/ha but presented significant risks to food security. Regional evaluation indicated that Central Africa has the greatest biogas potential (18,400 m³/ha at 16.8 t/ha yield), in contrast to East Africa's semi-arid region (9,200 m³/ha at 9.5 t/ha). Life-cycle GHG assessment revealed cassava bioenergy emissions of 30–40 gCO₂-eq/MJ, which are notably less than diesel (95 gCO₂-eq/MJ), gasoline (93 gCO₂-eq/MJ), and coal-generated power (110 gCO₂-eq/MJ). These results show ANN as a revolutionary resource for enhancing cassava bioenergy, reconciling food–energy conflicts, and advancing UN SDGs 7, 12, and 13 across Africa.\u003c/p\u003e","manuscriptTitle":"Optimizing Cassava-to-Bioenergy Conversion Using Artificial Neural Networks: A Sustainable Pathway for Africa","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-23 08:35:10","doi":"10.21203/rs.3.rs-7671514/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":"6254ff2e-f4df-4f5c-85fe-7fae86909cd0","owner":[],"postedDate":"September 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":55106258,"name":"Agricultural Engineering"},{"id":55106259,"name":"Artificial Intelligence and Machine Learning"},{"id":55106260,"name":"Food Science \u0026 Technology"}],"tags":[],"updatedAt":"2025-10-12T17:23:18+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-23 08:35:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7671514","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7671514","identity":"rs-7671514","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}