GRADIENT DECENT BOOSTED NEURAL NETWORK (GBNN) BASED DIABETES RISK PREDICTION WITH MINIMIZED ERROR LOSS USING RETINAL IMAGES

Abstract

Data mining is defined as the systematic approach used to extract processed information from user data. Data mining consists of different techniques such as Classification, Clustering and Regression can be used to solve real time problems. These techniques have number of algorithms that is only used under specific techniques but neural network is the miscellaneous algorithm can be used in Classification and Clustering techniques. Neural network can be implemented with different learning techniques such as Supervised, Unsupervised, Deep and transfer learning. Learning techniques used to learn the information from the data given. This work focuses on implementing neural network with gradient boosting to optimize neural network. The purpose of network optimization is to minimize the error. The proposed technique shows high accuracy of 90% with low error rate. The error rate of the proposed technique is compared existing neural network.