Efficient Machine Learning Models for the Accurate Prediction of Diabetes

Aim: Identifying and predicting the diabetes for a patient using Random Forest (RF), Naive Bayes (NB), and Multi- Layer Perceptron (MLP). Methods and Material: The SVM algorithm, known as group 1, is combined in the project to improve diabetes prediction with three other algorithms: Multi-Layer Perceptron, Random Forest and Naive Bayes, grouped as 2, 3, and 4. This multi-algorithmic framework seeks to improve efficiency and accuracy by utilising the robust classification of SVM, the ensemble learning of RF, the pattern capture of Multilayer Perceptron, and the probabilistic approach of NB. This comprehensive solution goes beyond the limitations of the conventional SVM-based model. Result: With the SVM, MLP, RF, and NB algorithms, our machine learning models showed noteworthy accuracy and promising results in the prediction of diabetes. RF improved accuracy by 76%, Naive Bayes by 78%, and MLP by 82%, according to the data. Conclusion: In diabetes prediction, combining SVM with RF, MLP, and NB shows a thorough approach, showing encouraging results and proving that MLP is effective for early diagnosis and treatment, outperforming conventional SVM-based models with improved accuracy.