International Journal of Hydrogen Energy, Год журнала: 2025, Номер 102, С. 1377 - 1398
Опубликована: Янв. 17, 2025
Язык: Английский
Процитировано
2Robust predictive framework for diabetes classification using optimized machine learning on imbalanced datasets
Frontiers in Artificial Intelligence, Год журнала: 2025, Номер 7
Опубликована: Янв. 7, 2025
Diabetes prediction using clinical datasets is crucial for medical data analysis. However, class imbalances, where non-diabetic cases dominate, can significantly affect machine learning model performance, leading to biased predictions and reduced generalization. A novel predictive framework employing cutting-edge algorithms advanced imbalance handling techniques was developed. The integrates feature engineering resampling strategies enhance accuracy. Rigorous testing conducted on three datasets-PIMA, Dataset 2019, BIT_2019-demonstrating the robustness adaptability of methodology across varying environments. experimental results highlight critical role selection mitigation in achieving reliable generalizable diabetes predictions. This study offers significant contributions informatics by proposing a robust data-driven that addresses challenges, thereby advancing
Язык: Английский
Процитировано
1Efficient diagnosis of diabetes mellitus using an improved ensemble method
Scientific Reports, Год журнала: 2025, Номер 15(1)
Опубликована: Янв. 25, 2025
Diabetes is a growing health concern in developing countries, causing considerable mortality rates. While machine learning (ML) approaches have been widely used to improve early detection and treatment, several studies shown low classification accuracies due overfitting, underfitting, data noise. This research employs parallel sequential ensemble ML paired with feature selection techniques boost accuracy. The Pima India Data from the UCI Repository served as dataset. preprocessing included cleaning dataset by replacing missing values column means selecting highly correlated features using forward backward methods. was split into two parts: training (70%), testing (30%). Python for Jupyter Notebook, there were design phases. first phase utilized J48, Classification Regression Tree (CART), Decision Stump (DS) create random forest model. second employed same algorithms alongside methods—XG Boost, AdaBoostM1, Gradient Boosting—using an average voting algorithm binary classification. Evaluation revealed that XG Boosting achieved of 100%, performance metrics including F1 score, MCC, Precision, Recall, AUC-ROC, AUC-PR all equal 1.00, indicating reliable predictions diabetes presence. Researchers practitioners can leverage predictive model developed this work make quick mellitus, which could save many lives.
Язык: Английский
Процитировано
1Stacking and ridge regression-based spectral ensemble preprocessing method and its application in near-infrared spectral analysis
Talanta, Год журнала: 2024, Номер 276, С. 126242 - 126242
Опубликована: Май 11, 2024
Язык: Английский
Процитировано
7Improved pediatric ICU mortality prediction for respiratory diseases: machine learning and data subdivision insights
Respiratory Research, Год журнала: 2024, Номер 25(1)
Опубликована: Май 23, 2024
Abstract The growing concern of pediatric mortality demands heightened preparedness in clinical settings, especially within intensive care units (ICUs). As respiratory-related admissions account for a substantial portion illnesses, there is pressing need to predict ICU these cases. This study based on data from 1188 patients, addresses this imperative using machine learning techniques and investigating different class balancing methods prediction. employs the publicly accessible “Paediatric Intensive Care database” train, validate, test model predicting patient mortality. Features were ranked three feature selection techniques, namely Random Forest, Extra Trees, XGBoost, resulting 16 critical features total 105 features. Ten models ensemble are used make accurate predictions. To tackle inherent imbalance dataset, we applied unique partitioning technique enhance model's alignment with distribution. CatBoost achieved an area under curve (AUC) 72.22%, while stacking yielded AUC 60.59% proposed subdivision technique, other hand, provides significant improvement performance metrics, 85.2% accuracy 89.32%. These findings emphasize potential enhancing prediction inform strategies improved readiness.
Язык: Английский
Процитировано
4Software Defect Prediction Based on a Multiclassifier with Hyperparameters: Future Work
Results in Engineering, Год журнала: 2025, Номер unknown, С. 104123 - 104123
Опубликована: Янв. 1, 2025
Язык: Английский
Процитировано
0Metaheuristic Feature Selection for Diabetes Prediction with P-G-S Approach
Procedia Computer Science, Год журнала: 2025, Номер 252, С. 165 - 171
Опубликована: Янв. 1, 2025
Язык: Английский
Процитировано
0An improved decision tree model through hyperparameter optimization using a modified gray wolf optimization for diabetes classification
Computer Methods in Biomechanics & Biomedical Engineering, Год журнала: 2025, Номер unknown, С. 1 - 17
Опубликована: Фев. 13, 2025
Diabetes is a chronic condition that affects blood sugar levels and vital organs in the body. Early detection crucial given increasing global prevalence of diabetes grave risk complications if not properly managed. Thus, good prediction system necessary. Although Decision Tree (DT) commonly used for classification, it less effective with large datasets. We propose hyperparameter optimization DT using Grey Wolf Optimization (GWO), which has exploration both exploitation capabilities. However, limited search space GWO may hinder practical exploitation, leading to premature optimization. To address this, we modified (MGWO) by adding Levy distribution function enhance movements alpha, beta, delta wolves. also provide GA (Genetic Algorithm) as comparative algorithm. The fitness value MGWO 0.8498, surpassing (0.8373) (0.8492). Evaluation results indicate yield similar superior accuracy compared GWO. proposed method outperforms existing ones. Further research needed evaluate impact varying number wolves on performance classification accuracy.
Язык: Английский
Процитировано
0An Analytical Predictive Model and Secure Wed Based Personalized Diabetes Monitoring System using Stacking Ensemble Classification
Deleted Journal, Год журнала: 2024, Номер 2(04), С. 967 - 975
Опубликована: Апрель 22, 2024
As the number of people diagnosed with diabetes continues to rise, this study takes a groundbreaking approach by developing secure web-based personalized monitoring system that incorporates analytical prediction models. This technology was developed in response critical need for sophisticated solutions address unique demands each patient. The suggested method aims transform treatment using predictive modeling anticipate diabetic trends and possible consequences. demonstrates strong commitment security creating platform handles patient data highest level care. By resolving important privacy problems trustworthy environment users, framework guarantees protection sensitive health information. use several models' characteristics make trend estimates more accurate reliable. In addition improving system's skills, stacking ensemble classification helps it adapt different profiles. Because importance accessibility usability encouraging participation, solution revolves around creation user-friendly online interface. interface is living, breathing allows frictionless communication between healthcare practitioners their patients. With help tailored insights predictions, patients are better able manage diabetes. At same time, doctors hospitals have access all information they need, which them take better, preventative measures.
Язык: Английский
Процитировано
0Clinical applications of artificial intelligence in diabetes management: A bibliometric analysis and comprehensive review
Informatics in Medicine Unlocked, Год журнала: 2024, Номер 50, С. 101567 - 101567
Опубликована: Янв. 1, 2024
Язык: Английский
Процитировано
0