Comprehensive Diabetes Prediction Applying Fused Machine Learning

Comprehensive Diabetes Prediction Applying Fused Machine Learning

B. Surya Sai Kiran Akash 1, D. Surya Chandra Varma, B. Manoj3, B.Harika4 , K. Pavan Kumar5

1 Department of Computer Science & Engineering: Raghu Engineering College

2 Department of Computer Science & Engineering: Raghu Engineering College

 3 Department of Computer Science & Engineering: Raghu Engineering College

4 Department of Computer Science & Engineering: Raghu Engineering College

5 Associate Professor, Department of Computer Science & Engineering: Raghu Engineering College

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - Early and accurate disease prediction is essential for effective prevention and management of medical conditions, particularly for globally prevalent diseases such as diabetes, which has become increasingly common due to modern dietary habits and sedentary lifestyles. This study introduces an advanced diabetes prediction model that leverages machine learning (ML) techniques to enhance diagnostic accuracy. The system integrates multiple classification algorithms, including Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbors (KNN), within a voting classifier framework, further enhanced by a fuzzy logic module to improve prediction performance. The model is trained using a standard American hospital dataset obtained from Kaggle, with 80% allocated for training and 20% for testing. Beyond disease detection, the system offers personalized recommendations on diet, physical activity, and routine health checkups by analyzing real-time medical records, ensuring a tailored approach to patient management. To facilitate scalability, security, and real-time accessibility, the system is deployed on a cloud platform, allowing seamless integration with healthcare applications. This cloud-based deployment ensures secure access for both healthcare professionals and patients from any internet-enabled device, enhancing usability while safeguarding sensitive medical information. Achieving 94% accuracy, the proposed fused ML model surpasses traditional prediction methods, and future enhancements aim to further refine the model, incorporate larger datasets, and expand its applicability to other diseases, demonstrating the transformative potential of ML and cloud-based healthcare analytics.

Key Words:  Support Vector Machine, K-Nearest Neighbor, Random Forest, Voting Classifier, Fuzzy Logic Module