Analysis of Diabetes Patients using Classification algorithms

摘要

Medical applications find classification algorithms for analyzing the patient’s condition. Classification algorithms in turn are of many types and deals with variety of characteristics of the clinical/patient data. To identify a good classification algorithm and achieve better accuracy, kappa coefficient is always a challenge. In this paper, a convolutional neural network for 2-way classification of diabetes data is used. Various training/testing combinations, normalizations setup and batch size, epochs are tried out to get better classification performance via simulation in python.