Diabetes Prediction System Using Machine Learning

Abstract

Diabetes is a chronic health condition that affects millions of people worldwide, leading to severe complications if left undiagnosed or untreated. Early detection plays a critical role in managing diabetes effectively and reducing long-term health risks. Traditional diagnostic methods, such as blood tests, are often invasive, time-consuming, and require medical supervision. This research presents a machine learning-based system for predicting diabetes using patient health data. The study utilizes the PIMA Indian Diabetes Dataset, applying Random Forest Classifier due to its ability to handle complex data structures and deliver reliable results. To enhance accuracy, the dataset undergoes preprocessing, including feature scaling with StandardScaler and handling missing values. The model is evaluated using multiple performance metrics, including accuracy, precision, recall, and F1-score, ensuring a comprehensive assessment of its effectiveness. Additionally, the system is deployed using Gradio, providing an intuitive interface where users can input key health indicators—such as glucose levels, BMI, blood pressure, and age—to receive an instant prediction.

Future enhancements include integrating deep learning techniques and expanding the system to work with real-time health monitoring devices. This research highlights the potential of artificial intelligence in non-invasive and early diabetes detection, contributing to more accessible and proactive healthcare solutions.