Enhanced Prediction of Diabetes Using Integrated Machine Learning and Deep Learning Techniques

Diabetes mellitus is still a considerable public health issue worldwide. Recent advances in machine learning (ML) and deep learning (DL) offer an exciting set of tools to enable early prediction and personalized treatment. This study evaluates a hybrid approach combining ML and DL algorithms to improve the predictive accuracy for diabetes. We implemented data preprocessing, feature extraction, and model tuning using the Pima Indians Diabetes dataset. Various algorithms, including Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors, SVM, XGBoost, and Artificial Neural Networks, were compared using accuracy, precision, recall, and F1-score. The performance evaluation in this study confirms that DL models outperform traditional ML techniques in recognizing complex patterns, which can help provide better early interventions, thereby improving clinical outcomes.