Diagnosis of TBC Disease Using SVM and Feedforward Backpropagation

Abstract

Tuberculosis (TBC) is an infectious disease caused by a virus Mycobacterium tuberculosis. One of the organs that is often infected by the virus Mycobacterium tuberculosis is the lungs. This disease is the second largest killer worldwide for infectious diseases after HIV/AIDS (Laily et al., 2015). Therefore, the level of diagnosis accuracy TBC disease needs to be improved using better methods. After the data is collected, then the data is processed in the preprocessing stage and through the normalization process so that the data range can be balanced. Furthermore, the last process is the classification process. In this classification process using two methods, namely Support Vector Machine and Feedforward Backpropagation. The two classification methods are assessed because they are simple and has a fairly precise level of accuracy. But also has a weakness in the selection of appropriate features. Based on research that has been done, using model testing with 10 executions, the accuracy results for Support Vector Machine produces an accuracy of 97.41%, while the results accuracy for Feedforward Backpropagation produces a level of accuracy by 98.51%. This shows that the Feedforward method Backpropagation is considered to improve the accuracy of diagnosis TBC disease.