Smart Medical Health Prediction Application Using Data Mining Integrated With Deep Learning for Cataract Detection

Abstract

Cataract is a major cause of avoidable vision loss, and early screening is essential to prevent blindness. This work proposes a smart medical health prediction system that combines data mining with deep learning to automatically detect cataract from ocular images. Preprocessing techniques are used to clean and enhance images, while a convolutional neural network classifies eyes as normal or cataract-affected. The system is designed as an assistive tool to provide fast, low-cost, and reliable screening support, especially in resource-limited settings. Experimental results show good accuracy and sensitivity, demonstrating that the integrated approach improves early detection and supports clinical decision- making in ophthalmology.