Abstract
Untreated diabetic retinopathy, a complication of long-term high blood sugar levels, may lead to total blindness if it is not caught and treated quickly. Thus, in order to avoid its devastating consequences, diabetic retinopathy must be medically diagnosed and treated early. Diabetic retinopathy is difficult to diagnose manually, thus patients often have to wait a long period before receiving treatment from an ophthalmologist. With the use of an automated technology, we can discover diabetic retinopathy early and begin treatment immediately to prevent additional damage to the eye. The present study proposes a machine learning strategy for extracting three features-exudates, haemorrhages, and micro aneurysms and classifying them with the help of a hybrid classifier comprised of components from the support vector machine, k nearest neighbour, random forest, logistic regression, and multilayer perceptron network.