An Initial Screening Method for Tuberculosis Diseases Using a Multi-objective Gradient Evolution-Based Support Vector Machine and C5.0 Decision Tree

Abstract

Tuberculosis is one of the top ten causes of death worldwide. Although this disease is curable and preventable, yet many new tuberculosis cases still occur especially in developing countries. Many low-income families cannot afford the medical diagnosis for tuberculosis. Therefore, this paper proposes an initial screening for tuberculosis infection using a data mining approach. In this paper, the initial screening is conducted using classification method developed from non-linear support vector machine and gradient evolution algorithm. Herein, the gradient evolution algorithm is able to find the best parameter setting for the support vector machine algorithm. The classification is performed based on some information which can be easily collected without medical test. The proposed algorithm is also compared with some other metaheuristic-based support vector machine algorithms. The experimental results show that the proposed algorithm has promising results shown by the small error rate. In addition, a C5.0 decision tree is employed to further analyze the rules of TB infection. The result reveals that people who are protected with vaccine still possible to be infected by tuberculosis, especially if there is direct contact with an active tuberculosis patient. Furthermore, people with low body mass index and low education has higher risk to get tuberculosis infection. The result of this study could help people conduct self-diagnosis for the tuberculosis infection before deciding to do a medical test.