CNN Architecture for Diabetes Classification

Abstract

Machine Learning (ML) algorithms deal with only linear data and here comes the necessity of deep learning (DL) algorithms to deal with non-linear input datas. Also, memory requirements and computational costs are addressed by DL algorithms used for classification for medical applications. In this paper, we have used convolutional neural network (CNN) which is the basic deep learning algorithm for 2-way diabetes classification. It helps in attaining better accuracy, kappa coefficient which are premier metrics to decide performance of classification. Simulations are carried out using python.