Performance Evaluation of Naive Bayes and Back Propagation Neural Network classifiers in gestational Diabetes Mellitus Classification

Abstract

The classification of Gestational Diabetes plays significant role in healthcare domain. Gestational Diabetes Mellitus (GDM) is a condition in which a hormone made by the placenta prevents the body from using insulin effectively. Glucose builds up in the blood instead of being absorbed by the cells.GDM is getting more common as overweight as well as next medical disorders are growing. It needs diligent and knowledgeable medical supervision, because the treatment provided throughout pregnancy impacts not only obstetric safety but also the likelihood of growing type 2 disease many years into the further, in both the infant and the mother. In this paper, precision of GDM classification is determined by comparing Back Propagation Neural Network (BPNN) and Enhanced Naïve Bayes classifiers. BPNN classifier is a type of ANN classifier that is developed for classification of GDM disease with accuracy estimation. The Enhanced Naïve Bayes classifier is the ML classification technique that is developed. The classification assessment methods Confusion Matrix (CM) and Balanced Error Rate(BER) are used to compare both the classification methodologies' efficiency and accuracy. The experimental results show that proposed Enhanced Navie Bayes methodology is having higher degree of efficiency and accuracy.