Predicting the existence of mycobacterium tuberculosis infection by Bayesian Networks and Rough Sets

Abstract

A correct diagnosis of tuberculosis can be only stated by applying a medical test to patient's phlegm. The result of this test is obtained after a time period of about 45 days. The purpose of this study is to develop a data mining solution which makes diagnosis of tuberculosis as accurate as possible and helps deciding if it is reasonable to start tuberculosis treatment on suspected patients without waiting the exact medical test results. In this research, we compared the use of Bayesian Networks and Rough Sets to predict the existence of mycobacterium tuberculosis. 503 different patient records having 30 separate input parameters are obtained from a private clinic and used in the entire process of this research. The Bayesian Network model classifies the instances with RMSE of 22% whereas Rough Set algorithm does the same classification with RMSE of 37%. As a result, Bayesian Network is an accurate and reliable method when compared with Rough Set method for classification of tuberculosis patients.