Prediction of Diabetic Retinal Diseases Using Random Forest Machine Learning Algorithm
Abstract
Diabetes is the major widespread disease among human beings and hotspot in medical science. This escorts the rigorous long-term diabetic retinal diseases without any indications. The record of chronicle condition provides vision loss. Early detection of diabetic retinal diseases is obligatory to avoid vision challenges in timing treatment. Auspiciously, the diabetic patient’s outcome has to predict the risk of diabetic retinal diseases like Diabetic Retinopathy, Glaucoma, and Diabetic Macular Edema. On the support of Machine learning algorithms, near the beginning of the diseases can foretell. Applied Support Vector Machine (SVM), Logistic regression, boosted logistic, k-nearest neighbor (k-nn) and Random Forest (RF) to formulate a final model. Based on the accuracy the random forest acted as a final model for testing the data to predict the diseases. The accuracy of 0.98 obtained in the training model and 0.94 of accuracy reached in the testing with 0.93, 0.95, and 0.96 of sensitivity, specificity, and precision. This consequence takes a vital role in the clinical expert system to make decisions in a moment.