Machine Learning Algorithm Applied for Predicting the Presence of Mycobacterium Tuberculosis

Abstract

Tuberculosis is a contagious disease that causes of death. The body's response to active TB infection produces inflammation that can damage the lungs. Areas affected by active TB gradually fill with scar tissue. It is spread from person-to-person. A person is often infected by inhaling the germs. Tuberculosis germs are spread into the air when a person with TB disease of the lungs or throat coughs, sneezes, speaks, or sings. These germs can stay in the air for several hours, depending on the environment. However, patients with chronic diseases, such as diabetes, chronic kidney disease, and silicosis, are at elevated risk. Finally, age younger than 4 years, long-term malnutrition, and substance abuse are independent risk factors for disease. The aim of this research work is to develop an adaptive Neuro-Fuzzy system for predicting the presence of Mycobacterium tuberculosis. The system is structured with inputs and one output of which rules were generated by the system with the help of three domain Medical expertise and are injected in to the knowledge based where the system would use this rules to make decisions and draw a conclusion. MATLAB7.0 was used to implement this experiment using fuzzy logic and Neural Network toolbox. In this experiment linguistic variables are evaluated using Gaussian membership function.