Optimal Model Parameters Research Articles

ICESat-2 data denoising and forest canopy height estimation using Machine Learning

Supervised classification methods can distinguish between noise and signal in ice, cloud, and land elevation satellite-2 (ICESat-2) data across various feature perspectives and autonomously optimize parameters. Nevertheless, model generalization remains a significant limitation for practical applications. This study focuses on developing a universal denoising model for ICESat-2 using machine learning algorithms and analyzing its spatial transferability under various forest and terrain conditions. A photon-denoising feature parameter system is developed based on the analysis of the three-dimensional distribution of photons in forested regions. This system reduces the parameters dependent on absolute physical quantities and increases those that are less influenced by terrain and forest features to enhance the model’s transferability. Subsequently, automated machine learning algorithms (AutoML) are used for model selection and parameter optimization across six non-parametric regression models. We evaluate the accuracies of the local, global, and transfer models in estimating canopy height across four representative forested areas in China. Results show that the algorithm can effectively distinguish between signal and noise photons. The estimated canopy heights from signal photons are highly consistent with heights obtained using airborne laser scanning (ALS), exhibiting a Pearson correlation coefficient (r) of 0.89, root mean square errors (RMSE) of 3.75 m, relative root mean square error (rRMSE) of 0.27, relative bias (rBias) of −0.11 and mean Bias of −1.45 m. Notably, the accuracy of canopy height estimation by the global model has increased by an average of 21 % compared to ICESat-2 land-vegetation along-track products (ATL08). Furthermore, the model exhibits significant spatial transfer capabilities, with the accuracies of the transfer model exceeding those of ATL08 by margins ranging from 4 % to 41 %. This study marks a significant advancement in photon-denoising methodologies, providing a robust and transferable solution for large-scale environmental data analysis.

ENSG: Enhancing Negative Sampling in Graph Convolutional Networks for Recommendation Systems

In the field of recommendation, negative samples that are close to positive samples are referred to as “hard negative samples”. These hard negative samples are more likely to be incorrectly recommended to users. Therefore, researching how to enhance the ability to identify hard negative samples and thereby improve recommendation accuracy is an important direction in the field of recommendation systems. To address this issue, we propose a new model named Enhancing Negative Sampling in Graph Convolutional Networks for Recommendation Systems (ENSG). Firstly, part of the information in the positive sample was injected into the negative sample, and the high-quality, difficult negative sample was selected by the inner product method. Secondly, a lightweight graph convolutional network is used for feature extraction to obtain node representations for both users and items. Then, contrastive learning is used to maximize the similarity between the hard negative samples and the positive samples, so that the hard negative samples are as close to the positive samples as possible in the feature space. Ultimately, the model is trained by taking hard negative samples as input and minimizing the loss function to optimize model parameters; the recommendation accuracy of the model is enhanced. To validate the effectiveness of our approach, we conducted comparative evaluations utilizing three publicly accessible datasets. The experimental outcomes indicate that our proposed model surpasses the baseline model in terms of performance. On three public datasets, the algorithm proposed in this paper achieved the highest improvements of 16.27% and 12.72% in the evaluation metrics Recall@20 and NDCG@20, respectively.

Innovations in animal health: artificial intelligence-enhanced hematocrit analysis for rapid anemia detection in small ruminants

Due to their value as a food source, fiber, and other products globally, there has been a growing focus on the wellbeing and health of small ruminants, particularly in relation to anemia induced by blood-feeding gastrointestinal parasites like Haemonchus contortus. The objective of this study was to assess the packed cell volume (PCV) levels in blood samples from small ruminants, specifically goats, and create an efficient biosensor for more convenient, yet accurate detection of anemia for on-farm use in agricultural environments for animal production optimization. The study encompassed 75 adult male Spanish goats, which underwent PCV testing to ascertain their PCV ranges and their association with anemic conditions. Using artificial intelligence-powered machine learning algorithms, an advanced, easy-to-use sensor was developed for rapidly alerting farmers as to low red blood cell count of their animals in this way to enable timely medical intervention. The developed sensor utilizes a semi-invasive technique that requires only a small blood sample. More precisely, a volume of 30 μL of blood was placed onto Whatman filter paper No. 1, previously soaked with anhydrous glycerol. The blood dispersion pattern on the glycerol-infused paper was then recorded using a smartphone after 180 s. Subsequently, these images were examined in correlation with established PCV values obtained from conventional PCV analysis. Four separate machine learning models (ML) supported models, namely support vector machine (SVM), K-nearest neighbors (KNN), backpropagation neural network (BPNN), and image classification-based Keras model, were created and assessed using the image dataset. The dataset consisted of 1,054 images that were divided into training, testing, and validation sets in a 70:20:10 ratio. The initial findings indicated a detection accuracy of 76.06% after only 10 epochs for recognizing different levels of PCV in relation to anemia, ranging from healthy to severely anemic. This testing accuracy increased markedly, to 95.8% after 100 epochs and other model parameter optimization. Results for SVM had an overall F1 score of 74–100% in identifying the PCV range for blood pattern images representing healthy to severely anemic animals, and BPNN showed 91–100% accuracy in identifying the PCV range for anemia detection. This work demonstrates that AI-driven biosensors can be used for on-site rapid anemia detection. Optimized machine learning models maximize detection accuracy, proving the sensor’s validity and rapidity in assessing anemia levels. This breakthrough will allow farmers, with rapid results, to increase animal wellbeing and agricultural productivity.

Spatial Spillover Effects of Smallholder Households’ Adoption Behaviour of Soil Management Practices Among Push–Pull Farmers in Rwanda

Push–pull technology (PPT) integrates maize with the legume fodder Desmodium sp. and the border crop Brachiaria sp., aiming to enhance maize production in Rwanda. Despite its potential, the adoption of complementary soil management practices (SMP), vital for PPT’s success, remains low. This study employs spatial econometric methods to evaluate the determinants of SMP adoption and the interdependencies in decision-making among PPT-practicing farmers. We constructed a spatial weight matrix based on a global Moran’s I index and identified optimal model parameters through principal component analysis. Utilizing a spatial Durbin probit model (SDPM), we assessed the spatial interdependence of SMP adoption decisions among maize farmers. Our findings reveal significant spatial dependence in SMP adoption within a 1.962 km radius, with improved seed usage, household income, yield, farmer group membership and size of land cultivated being key factors positively influencing adoption. We propose a “nonequilibrium promotion strategy” to enhance SMP adoption, emphasizing the establishment of pilot regions to broaden outreach. Additionally, fostering technical training and selecting farmers with adequate resources as demonstration leaders can enhance spatial spillover effects. This research provides insights for developing policies to scale up push–pull technology in Rwanda and across Sub-Saharan Africa.

The improved de Bruijn graph for multitask learning: predicting functions, subcellular localization, and interactions of noncoding RNAs.

Noncoding RNA refers to RNA that does not encode proteins. The lncRNA and miRNA it contains play crucial regulatory roles in organisms, and their aberrant expression is closely related to various diseases. Traditional experimental methods for validating the interactions of these RNAs have limitations, and existing prediction models exhibit relatively limited functionality, relying on isolated feature extraction and performing poorly in handling various types of small sample tasks. This paper proposes an improved de Bruijn graph that can inject RNA structural information into the graph while preserving sequence information. Furthermore, the improved de Bruijn graph enables graph neural networks to learn broader dependencies and correlations among data by introducing richer edge relationships. Meanwhile, the multitask learning model, DVMnet, proposed in this paper can handle multiple related tasks, and we optimize model parameters by integrating the total loss of three tasks. This enables multitask prediction of RNA interactions, disease associations, and subcellular localization. Compared with the best existing models in this field, DVMnet has achieved the best performance with a 3% improvement in the area under the curve value and demonstrates robust results in predicting diseases and subcellular localization. The improved de Bruijn graph is also applicable to various scenarios and can unify the sequence and structural information of various nucleic acids into a single graph.

Real-Time prediction of pool fire burning rates under complex heat transfer effects influenced by ullage height: A comparative study of BPNN and SVR

This research utilizes machine learning methods to forecast the complex, non-linear thermal phenomena, along with heat transfer mechanisms, that influence the burning rate of pool fires, especially with changes in ullage height. Experiments involving pool fires were systematically designed and carried out, incorporating different diameters and ullage heights. Heptane was used as the representative alkane fuels. A dataset containing more than 70,000 sets of data was created as a training dataset for training the Backpropagation Neural Network (BPNN) and Support Vector Regression (SVR) models. During the optimization of machine learning model parameters, this study is based on Particle Swarm Optimization (PSO) with the principle of intelligent optimization to efficiently and accurately screen and optimize the key parameters of the model. The combustion duration, pool dimensions, and non-dimensional ullage height were input into a machine-learning model to predict the burning rate. By comparing against experimental data, the model was found to be able to predict the dynamic evolution of the burning rate of the pool fire in a real-time manner. The SVR model demonstrates greater predictive accuracy in comparison to the BPNN model, and the relative prediction error remains within ± 20 %, which fully proves its effectiveness and generalization ability in the prediction of pool fire burning rate. The insights gained will offer substantial scientific backing for enhanced fire monitoring systems, while highlighting the capability of advanced machine learning methodologies to predict the intricate, real-time thermal dynamics and heat transfer characteristics of burning liquid fuels.

Influence of ambient temperature and humidity on the evolution of the tension softening behavior in concrete

The tension softening relationship is a crucial input parameter for modeling the crack propagation in concrete. To examine the influence of curing conditions on this parameter, fracture tests were performed on specimens kept at different temperatures and humidity levels across various ages. A best-fit maturity method for identifying optimal maturity model parameters was proposed and validated using test data. The results indicated that the area beneath the first branch of the softening curve increases, whereas the area under the tail decreases with ongoing hydration and higher humidity. The tensile strength ft and centroid coordinate σcg (vertical coordinate of softening curve area center) increase with rising temperatures up to 14 days, after which the trend reverses, with higher later-age ft and σcg observed in specimens exposed to elevated temperatures. Under high temperatures, the critical crack width wc (ultimate width at zero cohesive stress) and centroid coordinate wcg decrease. At the early stage, humidity has minimal effect on concrete softening properties. However, as the hydration progresses, the impact of decreasing humidity becomes more significant. At 60 days, ft, σcg, and wc are reduced by 23.3 %, 22.8 %, and 19.6 %, respectively, as humidity drops from 98 % to 30 %. The developed best-fit maturity method can predict various concrete properties with a maximum relative error within 7 %. The optimal activation energy Ea and moisture diffusion coefficient γ vary across temperature and humidity ranges, making it unfeasible to use a single set (Ea, γ) to predict properties under varying curing conditions.

Research on National Short-Term Precipitation Forecast Method Based on Machine Learning

Abstract. Climate change is closely related to human lives. With the development of information technology, meteorological data gradually show the characteristics of big data. The purpose of this study is to compare the application effects of different machine learning models in short-term precipitation prediction, to improve the accuracy and efficiency of prediction. Based on the annual average precipitation data of prefecture-level cities in China from 1990 to 2022, this paper uses Random Forest, eXtreme Gradient Boosting and Neural Networks to construct prediction models, and comprehensively evaluate them. In this study, the data is preprocessed to ensure the data quality of the input model. Next, the data is fed into three machine learning algorithm models, Random Forest, eXtreme Gradient Boosting and Neural Networks. Finally, the prediction performance of each model is evaluated by various indexes. In this paper, it is found that the Random Forest model has the best performance, and its prediction accuracy is higher than the other two models, which has great application potential in the field of short-term precipitation prediction. This study shows that a reasonable selection of machine learning methods and optimization of model parameters can effectively improve the accuracy of short-term precipitation prediction. This paper provides some empirical evidence for precipitation prediction, which will help to make more effective decisions in dealing with extreme weather events and climate change challenges in the future.

Research on National Short-Term Precipitation Forecast Method Based on Machine Learning

Abstract. Climate change is closely related to human lives. With the development of information technology, meteorological data gradually show the characteristics of big data. The purpose of this study is to compare the application effects of different machine learning models in short-term precipitation prediction, to improve the accuracy and efficiency of prediction. Based on the annual average precipitation data of prefecture-level cities in China from 1990 to 2022, this paper uses Random Forest, eXtreme Gradient Boosting and Neural Networks to construct prediction models, and comprehensively evaluate them. In this study, the data is preprocessed to ensure the data quality of the input model. Next, the data is fed into three machine learning algorithm models, Random Forest, eXtreme Gradient Boosting and Neural Networks. Finally, the prediction performance of each model is evaluated by various indexes. In this paper, it is found that the Random Forest model has the best performance, and its prediction accuracy is higher than the other two models, which has great application potential in the field of short-term precipitation prediction. This study shows that a reasonable selection of machine learning methods and optimization of model parameters can effectively improve the accuracy of short-term precipitation prediction. This paper provides some empirical evidence for precipitation prediction, which will help to make more effective decisions in dealing with extreme weather events and climate change challenges in the future.

Key phase diagram experiments and thermodynamic modeling of the ZnO‒B2O3, ZnO‒SiO2, and ZnO‒B2O3‒SiO2 systems

AbstractThermodynamic modeling of binary ZnO‒B2O3 and ZnO‒SiO2 systems and ternary ZnO‒B2O3‒SiO2 system were carried out based on the critical evaluation and optimization of all available phase diagram and thermodynamic property data using the CALculation of PHAse Diagram (CALPHAD) method. To resolve the inconsistency among the available phase diagram data in ternary ZnO‒B2O3‒SiO2 system, key phase diagram experiment was carried out using classical quenching method followed by electron probe micro‐analysis and X‐ray diffraction phase determination. The liquid oxide solution was described using the modified quasichemical model accounting for a short‐range ordering behavior in molten oxide solution. Thermodynamic models with optimized model parameters were used to predict the phase diagram and thermodynamic properties of the ternary systems.

Algorithm for simulating calligrapher’s stroke features using neural networks

In the field of calligraphy art, simulating the stroke features of calligraphers has always been a challenging task. Traditional methods often rely on manually designed rules or feature extraction algorithms, which are difficult to accurately capture calligraphy details and time-consuming. This article aimed to explore a more effective method to simulate the stroke features of calligraphers using neural network technology. This article mainly explored the simulation algorithm of calligrapher stroke features based on neural networks, including calligrapher stroke feature statistics, calligrapher stroke feature abstraction, and calligrapher stroke feature extraction. By using Convolutional Neural Network (CNN) to learn and analyze calligraphy works, a better performance neural network model was established to achieve automatic recognition and classification of calligrapher stroke features. By continuously optimizing model parameters, the accuracy of calligrapher stroke feature simulation can be improved. The average similarity rates for imitating stroke features of five calligraphers (regular script, cursive script, clerical script, seal script, and running script) using Artificial Neural Network (ANN) were 0.72, 0.62, 0.40, 0.33, and 0.53, respectively. The average similarity rates for imitating stroke features of 5 calligraphers (regular script, cursive script, clerical script, seal script, and running script) using CNN were 0.93, 0.78, 0.87, 0.67, and 0.80, respectively. The research results of this article promoted the inheritance and innovation of calligraphy art, and expanded the expression forms and application fields of calligraphy art.

Enhancing Lung Cancer Screening with Bidirectional LSTM and GRU Models

Abstract. This study aims to enhance lung cancer patient screening by developing and evaluating bidirectional Long Short-Term Memory (LSTM) and bidirectional Gated Recurrent Unit (GRU) models using the Lung Cancer dataset from Kaggle. The dataset includes 16 features related to patient symptoms and lung cancer status, providing a broad spectrum of symptoms to improve model accuracy. The research advances Artificial Intelligence (AI)-driven healthcare by integrating these sophisticated machine learning techniques into diagnostic processes. The methodology involves four main steps: preprocessing the dataset for model compatibility, defining the model architecture with bidirectional LSTM and GRU layers and evaluating its performance. The results show an overall accuracy of 52.17%, with accuracy, recall, and F1 scores for both cancerous and non-cancerous categories around 50%. Despite the hybrid model's average performance, it establishes a basis for future enhancements. Optimizing model parameters and exploring additional other techniques to improve prediction accuracy and clinical applicability will be done in the future.

Distribution of Pb, Zn, Fe, As, Sn, Sb, Bi, and Ni Between Oxide Liquid and Metal in the ‘CuO0.5’-CaO-AlO1.5 System in Equilibrium with Cu Metal at 1400 °C

AbstractThe increasing complexity of ore resources and recycled materials in the feed of pyrometallurgical processes present a technical challenge to the metallurgical engineers working on maximizing the recovery of the valuable elements and minimizing the environmental impact of the processes. To address this challenge, the availability of computational tools that can predict the mass and energy balance in complex systems is required. Then, the accurate description of phase equilibria in the complex multicomponent systems describing the chemistry of the pyrometallurgical processes becomes critical for the correct implementation of the indicated tools and facing the outlined industrial challenges. In the present study, the distribution of selected elements (Pb, Zn, Fe, As, Sn, Sb, Bi, and Ni) between oxide liquid and metal in the ‘CuO0.5’–CaO–AlO1.5 system in equilibrium with Cu metal at 1400 °C (liquidus of CaAl2O4) was experimentally studied using the equilibration and quenching technique followed by the electron probe X-ray microanalysis of the resulted samples. The study covered a wide range of effective p(O2) over the system from 10−11 to 10−3.5 (corresponding to formation of immiscible CuO0.5-rich slag). To avoid loss of volatile elements (Pb, Zn, As, Sn, Sb, and Bi), a correlation between ‘CuO0.5’ in oxide liquid and p(O2) in open system was obtained first, followed by studying the volatile elements distribution in closed conditions (Al2O3 crucible sealed in SiO2 ampoule), where ‘CuO0.5’ concentration was used as a marker to evaluate the effective p(O2) over the system. The experimental results were then used for the optimization of the thermodynamic model parameters of the system as part of the integrated experimental and self-consisting thermodynamic modeling research program of phase equilibria in the Cu–Pb–Zn–Fe–Ca–Si–Al–Mg–O–S–(As, Sn, Sb, Bi, Ag, Au, Ni, Cr, Co, and Na) gas/oxide liquid/matte/speiss/metal/solids system. Graphical Abstract

Enhanced predictability and interpretability of COVID-19 severity based on SARS-CoV-2 genomic diversity: a comprehensive study encompassing four years of data.

Despite the end of the global Coronavirus Disease 2019 (COVID-19) pandemic, the risk factors for COVID-19 severity continue to be a pivotal area of research. Specifically, studying the impact of the genomic diversity of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) on COVID-19 severity is crucial for predicting severe outcomes. Therefore, this study aimed to investigate the impact of the SARS-CoV-2 genome sequence, genotype, patient age, gender, and vaccination status on the severity of COVID-19, and to develop accurate and robust prediction models. The training set (n = 12,038), primary testing set (n = 4,006), and secondary testing set (n = 2,845) consist of SARS-CoV-2 genome sequences with patient information, which were obtained from Global Initiative on Sharing all Individual Data (GISAID) spanning over four years. Four machine learning methods were employed to construct prediction models. By extracting SARS-CoV-2 genomic features, optimizing model parameters, and integrating models, this study improved the prediction accuracy. Furthermore, Shapley Additive exPlanes (SHAP) was applied to analyze the interpretability of the model and to identify risk factors, providing insights for the management of severe cases. The proposed ensemble model achieved an F-score of 88.842% and an Area Under the Curve (AUC) of 0.956 on the global testing dataset. In addition to factors such as patient age, gender, and vaccination status, over 40 amino acid site mutation characteristics were identified to have a significant impact on the severity of COVID-19. This work has the potential to facilitate the early identification of COVID-19 patients with high risks of severe illness, thus effectively reducing the rates of severe cases and mortality.

Functional connectotomy of a whole-brain model reveals tumor-induced alterations to neuronal dynamics in glioma patients

Abstract Brain tumors can induce pathological changes in neuronal dynamics that are reflected in functional connectivity measures. Here, we use a whole-brain modeling approach to investigate pathological alterations to neuronal activity in glioma patients. By fitting a Hopf whole-brain model to empirical functional connectivity, we investigate glioma-induced changes in optimal model parameters. We observe considerable differences in neuronal dynamics between glioma patients and healthy controls, both on an individual and population-based level. In particular, model parameter estimation suggests that local tumor pathology causes changes in brain dynamics by increasing the influence of interregional interactions on global neuronal activity. Our approach demonstrates that whole-brain models provide valuable insights for understanding glioma-associated alterations in functional connectivity.

Enhancing Photovoltaic Solar Model Parameter Optimization: WSO-MTBO Hybrid Approach based on Newton-Raphson Method

Enhancing Photovoltaic Solar Model Parameter Optimization: WSO-MTBO Hybrid Approach based on Newton-Raphson Method

Characterizing and Understanding HGNN Training on GPUs

Owing to their remarkable representation capabilities for heterogeneous graph data, Heterogeneous Graph Neural Networks (HGNNs) have been widely adopted in many critical real-world domains such as recommendation systems and medical analysis. Prior to their practical application, identifying the optimal HGNN model parameters tailored to specific tasks through extensive training is a time-consuming and costly process. To enhance the efficiency of HGNN training, it is essential to characterize and analyze the execution semantics and patterns within the training process to identify performance bottlenecks. In this study, we conduct a comprehensive quantification and in-depth analysis of two mainstream HGNN training scenarios, including single-GPU and multi-GPU distributed training. Based on the characterization results, we reveal the performance bottlenecks and their underlying causes in different HGNN training scenarios and propose optimization guidelines from both software and hardware perspectives.

Predicting fungal infection sensitivity of sepals in harvested tomatoes using imaging spectroscopy and partial least squares discriminant analysis

Tomatoes ( Solanum lycopersicum L.) are a widely grown and globally traded vegetable, essential for both local consumption and international trade. However, approximately 30% of harvested tomato yields are lost due to fungal decay during postharvest handling. Timely disease identification is crucial to prevent such losses, but certain tomato varieties exhibit higher susceptibility to fungal infections than others. Additionally, there are variations in susceptibility among individual sepals, with unknown underlying causes. Traditional methods for assessing fungal presence in plants have limitations, such as sample destruction and a focus on symptom detection rather than evaluating susceptibility to fungal infection. Hence, there is a demand need for an accurate, non-destructive method capable of predicting susceptibility to fungal infection. The use of hyperspectral imaging (HSI) with chemometrics presents a pioneering approach to address this need. In this study, three tomato cultivars (‘Brioso,’ ‘Cappricia,’ and ‘Provine’) were studied. Hyperspectral images were captured on day-1 of harvest, followed by controlled fungal growth conditions. Ground truth assessments were conducted by three experts on day-3 and day-4, averaging severity scores assigned per sepal. The methodology involved extracting spectra from HSI images and calibrating and validating models using partial least squares discriminant analysis (PLSDA), aiming to optimize model parameters for accurate predictions. The models were categorized into those developed using data from a single variety (intravariety) and those utilizing data from multiple varieties combined (global models). The best-performing intravariety model was established using the Cappricia variety, achieving a balanced accuracy of 0.84. Conversely, a global model combining Cappricia and Provine varieties achieved a balanced accuracy of 0.70. Overall, the results suggest that distinguishing between more and less susceptible sepals is feasible under controlled conditions.

Design and optimization of a planar anti-buckling compliant rotational joint with a remote center of motion

Compliant mechanisms (CMs) exhibit some excellent mechanical properties and provide numerous innovative solutions for many existing mechanical applications. Among them, planar compliant rotational joints have made substantial contributions to precision engineering. To achieve high-precision positioning with a compliant rotational joint, it is essential to study characteristics such as anti-buckling and constant stiffness. The remote center of motion (RCM) mechanism, with its compact structure and ease of precise control, is expected to enhance overall rigidity and reduce parasitic displacements. Here, we propose a planar anti-buckling compliant rotational joint with a RCM. Static modeling and model validation are conducted for different geometric configurations, including single and combined structures. A distributed configuration with bidirectional anti-buckling properties is selected for optimization. Using a global parameter optimization model, single-objective optimization studies are conducted for three distinct characteristics. Subsequently, a multi-objective optimization model for constant stiffness and high precision is established. Based on the optimization results, a compliant rotational joint with bidirectional anti-buckling, constant rotational stiffness, and high precision is designed. Finally, the effectiveness of the optimization method is validated through physical experiments.

Physiological model-based machine learning for classifying patients with binge-eating disorder (BED) from the Oral Glucose Tolerance Test (OGTT) curve

Background and objective:Binge eating disorder (BED) is the most frequent eating disorder, often confused with obesity, with which it shares several characteristics. Early identification could enable targeted therapeutic interventions. In this study, we propose a hybrid pipeline that, starting from plasma glucose data acquired during the Oral Glucose Tolerance Test (OGTT), allows us to classify the two types of patients through computational modeling and artificial intelligence. Methods:The proposed hybrid pipeline integrates a classical mechanistic model of delayed differential equations (DDE) that describes glucose–insulin dynamics with machine learning (ML) methods. Ad hoc techniques, including structural identifiability analysis, have been employed for refining and evaluating the mathematical model. Additionally, a dedicated pipeline for identifying and optimizing model parameters has been applied to obtain reliable estimates. Robust feature extraction and classifier selection processes were developed to ensure the optimal choice of the best-performing classifier. Results:By leveraging parameters estimated from the mechanistic model alongside easily obtainable patient information (such as glucose levels at 30 and 120 min post-OGTT, glycated hemoglobin (Hb1Ac), body mass index (BMI), and waist circumference), our approach facilitates accurate classification of patients, enabling tailored therapeutic interventions. Conclusion:Initial findings, focusing on correctly categorizing patients with BED based on metabolic data, demonstrate promising outcomes. These results suggest significant potential for refinement, including exploration of alternative mechanistic models and machine learning algorithms, to enhance classification accuracy and therapeutic strategies.