Diabetic Retinopathy Risk Prediction for Diabetics Using Nearest Neighbour Approach

摘要

Non-communicable diseases (NCDs), in our study, include diabetes, hypertension and obesity, represent the global burden of diseases and cause deaths each year mainly in rural areas. Diabetic retinopathy (DR) is the prime root of irreversible blindness in the modern world. It refers to the retinal changes seen in patients with diabetes mellitus. In this paper, we present a novel approach to identify diabetic patients and predict the risk of diabetic retinopathy in diabetics using Naieve Bayes, z-ordering and kNN techniques. The results show that there is reduction in misclassification error and improvement in the efficiency of classification which help in improved screening of the dreaded complication of diabetes that aids in its early treatment.