Abstract
INTERNATIONAL JOURNAL OF ENGINEERING, SCIENCE AND - volume 14,issue 6,, June 2025
Pages: 81-95
OPTIMIZING ACCURACY OF DIABETES DISEASE DIAGNOSIS USING VOTING CLASSIFIER ALGORITHM
Varun Gupta, Dr. R.P.P Singh , Dr, Neeraj Marwaha
Category:Engineering, Science and Mathematics
Abstract:
Diabetes mellitus is a chronic metabolic disorder that poses a major threat to global health due to its increasing prevalence and potential for serious complications. Accurate and early diagnosis is essential to prevent disease progression and manage patient outcomes effectively. In this study, we propose the use of a Voting Classifier Algorithm, an ensemble learning approach that combines the predictive capabilities of multiple machine learning models, to improve the accuracy of diabetes diagnosis. The base models used include Logistic Regression, Random Forest, Support Vector Machine, and K-Nearest Neighbors. Both hard voting (majority class prediction) and soft voting (average of predicted probabilities) strategies are implemented to evaluate their effectiveness. The PIMA Indian Diabetes Dataset, which contains clinical data of female patients, is used for training and evaluation. Experimental results demonstrate that the Voting Classifier significantly outperforms individual classifiers in terms of accuracy, precision, recall, and F1-score. Notably, the soft voting strategy yields the highest diagnostic accuracy of over 97.82%, indicating the robustness of ensemble methods in medical prediction tasks. This study highlights the potential of Voting Classifiers in developing intelligent, data-driven healthcare solutions for early and reliable detection of diabetes, thereby aiding clinicians in informed decision-making.
Keywords: Machine Learning (ML), Diabetes Mellitus (DM), Healthcare System, Voting Classifier