Diabetes is a growing health concern worldwide, and early detection is crucial for effective treatment. This paper introduces a predictive model that combines Support Vector Machine (SVM) and Random Forest algorithms with a soft voting classifier to detect diabetes based on clinical data. The ensemble model leverages the strengths of both SVM and RF to improve overall performance. The methodology follows an Object Oriented Analysis and Design Methodology (OOADM) to model the system architecture and to address class imbalance issues, Synthetic Minority Over-sampling Technique (SMOTE) is employed as a data balancing technique. The proposed approach employs Electronic Health Records (EHR) and a user-friendly interface for efficient data analysis and real-time prediction incorporating attributes such as blood pressure, age, BMI, smoking status, and physical activity to predict diabetes. The results demonstrates the effectiveness of ensemble machine learning in identifying diabetes patients, achieving 96 % of accuracy, 97% of precision, 99% of recall, and 98% of F1-score highlighting its potential for early detection and providing a promising direction for future research and development of robust real-time prediction systems.
Goodness et al. (Fri,) studied this question.