<![CDATA[Tuberculosis (TBC or TB) is an infectious disease that usually attacks the lungs, caused by the bacterium Mycobacterium tuberculosis. WHO data report in 2006 put Indonesia as the third largest contributor of TB in the world. The high risk of dying of lung disease patients (18.7%) indicate that these diseases should be taken seriously. In addition to the lungs, where TB germs attack the brain and central nervous system, this will also lead to death (death). In this study, the author uses neuro-fuzzy system for diagnosing TB disease based mainly on clinical symptoms. Neuro-fuzzy systems are part of the major components forming soft computing, integrated between fuzzy systems and artificial neural networks. With the method of Adaptive Neuro-Fuzzy Inference System (ANFIS) in determining the classification rule with fuzzy logic that is able to provide a diagnosis like an expert whether someone is diagnosed: Negative TB, Other Disease and Positive TB. Based on ANFIS editor can be seen the results of measurements of the accuracy of the algorithm, the hybrid gets the same value of four types of membership function as Trapmf, gbellmf, gaussmf and psigmf of 99.99%. While the backpropagation algorithm produces different accuracies depending on each type of MF her. Where Trapmf membership type has an accuracy rate higher than the other three types of memberships by using the backpropagation algorithm. And to see what the diagnosis was designed using Matlab toolbox applications, such as appearance and surface at the FIS rule editor, diagnosis and therapeutic treatment.]]>
Copyrights © 2015
