Main Article Content
Diabetic Retinopathy (DR) is extracted under later stages of vision blur or complete blindness or vision loss. The diabetic retinopathy resultant causes an irreversible blindness and hence requires a detailed attention and early detection framework. In this article, we propose a machine learning approach for early segmentation and classification of diabetic retinopathy using pre-morbidities and patient history with respect to diabetics. Typically, a uncontrolled diabetic lead to retinopathy formation. The proposed technique classifies the possibilities of occurrences with feature and attributes of electronics health records (EHR) systems. The proposed technique includes multiple data-dimensions and attributes. Both segmentation and classification of attributes and feature-sets are cross validated via an inter-dependency mapping of attributes. The technique uses primary-dependencies attributes such as diabetics levels post and pre fasting, history of diabetic treatment and retinopathy aging with reference to aligned ill’ment. The technique has successfully classified with a performance ratio of 97.81% with respect to aimed parameter evaluation. The process has also retained a voice based monitoring and alerting system.