Predicting existence of Mycobacterium tuberculosis on patients using data mining approaches

A Correct diagnosis of tuberculosis (TB) 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(DM) 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 or not. In this research, we proposed the use of Sugeno-type “adaptive-network-based fuzzy inference system” (ANFIS) to predict the existence of mycobacterium tuberculosis. 667 different patient records which are obtained from a clinic are used in the entire process of this research. Each of the patient records consist of 30 separate input parameters. ANFIS model is generated by using 500 of those records. We also implemented a multilayer perceptron and PART model using the same data set. The ANFIS model classifies the instances with an RMSE of 18% whereas Multilayer Perceptron does the same classification with an RMSE of % 19 and PART algorithm with an RMSE of % 20. ANFIS is an accurate and reliable method when compared with Multilayer Perceptron and PART algorithms for classification of tuberculosis patients. This study has contribution on forecasting patients before the medical tests.