Intelligent System for Diagnosing Tuberculosis Using Adaptive Neuro-Fuzzy
Ibrahim Goni *
Department of Computer Science, Adamawa State University Mubi, Nigeria
Christopher U. Ngene
Department of Computer Engineering, University of Maiduguri, Nigeria
Manga I.
Department of Computer Science, Adamawa State University Mubi, Nigeria
Auwal Nata’ala
Department of Computer Science, The Federal Polytechnic Kaura Namoda, Zamfara State, Nigeria
Sunday J. Calvin
Department of Medical Laboratory Science, Adamawa State College of Health Technology Michika, Nigeria
*Author to whom correspondence should be addressed.
Abstract
Tuberculosis is a contiguous disease that is causing death both in developed and developing countries. The main aim of this research work was to a developed an intelligent system for diagnosing Tuberculosis using adaptive neuro-fuzzy methodology. Eleven symptoms of tuberculosis which are persistent cough for more than two weeks, cough with blood, weight loss, tiredness, chest pain, fever, difficulty in breathing, loss of appetite, lymph node enlargement, history of TB contact and night Sweat are assigned with weights which are categorize best on severity level as mild, moderate, severe and very severe, yes and no which serve as inputs to the adaptive neuro-fuzzy inference system (ANFIS). MATLAB 7.0 is used to implement this experiment, Trapezoidal Membership function was used, back propagation algorithm was used for training and testing, the error obtain is 0.41777 at epoch 2 which shows that the training performance is exactly 99.58223 and testing performance of the system are 99.58197 at epoch 2.
Keywords: Adaptive neuro-fuzzy, tuberculosis, Gaussian membership function, hybrid training algorithm, epoch