A Model for Prediction of Drug Resistant Tuberculosis Using Data Mining Technique

Abstract

The rate of mortality in the recent time because of tuberculosis disease is so alarming. Drug-Resistant Tuberculosis is a communicable disease very dangerous that attack lungs, many victims were not identified due to weak health systems facilities, poor doctor-patient relationship, and inefficient mechanisms for predicting of the disease. Data mining can be applied on medical data to foresee novel, useful and potential knowledge that can save a life, reduce treatment cost, increases diagnostic and prediction accuracy as well as delay taking during prediction which reduce the treatment cost of a patience. Several data mining technique such as classification, clustering, regression, and association rule were used to enhance the prediction of tuberculosis. In this project I used Naïve Bayes Classifier to design a model for predicting tuberculosis. I considered the following parameters; Gender, Chills, Fever, Night sweat, Fatigue, Cough with Blood, Weight loss, and Loss of Appetite for classification phase 1. While Gender Chest Pain, Sputum, Contact DR, Weight Loss, In-adequate treatment for classification phase 2 as the clinical symptom. The Naïve Bayes Classifier has the advantage of attribute independency, it is easy in construction, can classify categorical data, and can work on high dimensional data effectively. The model designed using Naïve Bayes Classifier is divided o into classification phase 1 and classification phase 2 and implemented using Phython 3.2 Programing Language. The result shows that Naïve Bayes Classfier was suitable in predicting drug resistant tuberculosis with performance accuracy of 82%, 98% and area under curve (AUC) is 88%.
 Keywords: Model Prediction, Tuberculosis. Drug, Resistant, Data Mining.