An Evolutionary Algorithm based Parameter Estimation using Pima Indians Diabetes Dataset

Abstract

Predictive modeling using the prowess of Machine Learning is getting stronger and smarter day by day. Often, these predictive models which are generally used to estimate specific values for a given problem are needed to be supplied with proper parameters. The parameters with which they are trained with have to be valuably optimal so that the models yield good results. In this paper, a neural Network is chosen as the predictive model in which Pima Indians Diabetes dataset is used. For obtaining the optimal values for the parameters of the Neural Network, Evolutionary Algorithm based approach has been used, which not only resulted in a better execution time but also generated more optimal values as compared to the other existing methods in terms of accuracy.