Analysis of Eye Disease Classification by Comparison of the Random Forest Method and K-Nearest Neighbor Method

Abstract

Eye disease is a serious issue all over the world, and image-based classification systems play an important role in the early detection and management of eye disease. This research compares the performance between Random Forest (RF) and K-Nearest Neighbor (KNN) classification models in identifying eye disorders using image datasets divided into four classes: "normal," "glaucoma," "cataract," and "diabetic retinopathy."   The dataset is converted into a feature vector and then divided into training data and test data subsets. The analysis results show that the RF model achieved an accuracy level of 80%, whereas the KNN model achieved 70%. Based on these findings, it is possible to conclude that the RF model outperforms the other models in categorizing the types of eye illnesses in the dataset. A Python-based website was also built utilizing the Flask framework to build an interactive and real-time eye illness diagnosis system. Users can upload photos of their retinas to this website and quickly receive eye disease detection results. The adoption of this technology has a tremendous impact, making eye disease detection solutions more accessible. Furthermore, this solution plays an important role in the early detection and effective management of eye health cases.