Experimental Analysis of Tuberculosis Classification Based on Clinical Data Using Machine Learning Techniques

The early detection of tuberculosis plays a significant rule to reduce the death rate of tuberculosis. However, the early detection of tuberculosis nowadays has a limitation such as it needs long periods of time to acquire accurate diagnosis because it includes many clinical examinations. To overcome this problem a new diagnosis schema is needed. This study evaluates the common machine learning techniques including Logistic Regression, K-Nearest Neighbour, Naive Bayes, Support Vector Machine, Random Forest, Neural Network and Linear Discriminant Analysis to diagnose tuberculosis using classification methods based on clinical data. The results show that most of machine learning techniques that use in this study have a good performance in classifying tuberculosis based clinical data. Those machine learning techniques have achieved 0.97-0.99 in testing F1-Score.