Diabetes Prediction: A Comprehensive Study Integrating Deep Learning and Machine Learning Approaches

Abstract

Diabetes, a prevalent chronic disease, affects a significant portion of the global population, with early detection and management being crucial for patient health. Despite the lack of a cure, advancements in machine learning (ML) and deep learning (DL) offer promising avenues for diabetes prediction. This study leverages the Pima Indian dataset to explore and compare the effectiveness of various ML and DL techniques, including Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), Random Forest, K-Nearest Neighbors (KNN), and Decision Tree. Our comprehensive evaluation, based on accuracy, precision, and recall, identified MLP as the most effective algorithm, achieving an accuracy rate of 85%. This research addresses current challenges in diabetes prediction by highlighting the superior performance of MLP in early detection, thereby underscoring its potential in improving diabetes management. The findings contribute to the ongoing discourse in healthcare technology, advocating for the integration of advanced ML techniques to enhance predictive accuracy and patient outcomes.