Identification of Tuberculosis with the Fuzzy Sugeno Method and Diet Recommendations Using the Naive Bayes Method
Eka Mala Sari Rochman, Retno Tri Lestari, Muhammad Ali Syakur, Hermawan Bin Fauzan. Kurniawan Eka Permana, Aeri Rachmad

University of Trunojoyo Madura

Abstract

The development of tuberculosis, according to the World Health Organization (WHO) in 2014, stated that it was estimated to affect 9.6 million people, with 12% of them being HIV-positive. Tuberculosis is a directly communicable disease caused by the tuberculosis bacterium (Mycobacterium Tuberculosis). Tuberculosis bacteria can be transmitted through physical contact, the air, the sputum of patients, and so on. Currently, many people are unaware of the early symptoms and dangers of tuberculosis, so an expert system is needed to diagnose tuberculosis early and provide dietary recommendations that can help expedite patient treatment. In this disease diagnosis expert system, the Fuzzy Sugeno method is used to make decisions by answering questions related to symptoms. Additionally, to ensure that patients have good nutritional status, dietary recommendations are provided using the Naive Bayes method to tailor diets according to the nutritional needs of tuberculosis patients. Using the Fuzzy Sugeno method for disease diagnosis resulted in an accuracy rate of 85.35%, and the Naive Bayes method for dietary habit recommendations produced the highest accuracy rate of 89.28% in k-fold = 3, with an average accuracy rate calculated across all folds of 78.57%.

Keywords: Mycobacterium Tuberculosis, Naive Bayes method, Fuzzy Sugeno, expert system

Topic: Artificial Intelligence and Data Science