Machine learning techniques for classification of diabetes and cardiovascular diseases

This paper presents the overview of machine learning techniques in classification of diabetes and cardiovascular diseases (CVD) using Artificial Neural Networks (ANNs) and Bayesian Networks (BNs). The comparative analysis was performed on selected papers that are published in the period from 2008 to 2017. The most commonly used type of ANN in selected papers is multilayer feedforward neural network with Levenberg-Marquardt learning algorithm. On the other hand, the most commonly used type of BN is Na'ive Bayesian network which shown the highest accuracy values for classification of diabetes and CVD, 99.51% and 97.92% retrospectively. Moreover, the calculation of mean accuracy of observed networks has shown better results using ANN, which indicates that higher possibility to obtain more accurate results in diabetes and/or CVD classification is when it is applied to ANN.