A Multi-layered Fuzzy Inference System for the Diagnosis of Hepatitis B

Abstract

Hepatitis B is a virus, which can attack the liver of a human body and leads to major disease of the liver. The prime aim or objective of this research work is to develop a medical diagnostic system to diagnose the Hepatitis B virus infection. A multi-layered fuzzy inference system has been proposed by using the Mamdani fuzzy model. In this proposed system, there are two layers, and both layers have different input variables. This system classifies the hepatitis B patient and non-hepatitis B patients into different classes. The input variables used for layer 1 are jaundice, dark urine, abdominal pain and vomiting. Similarly, layer 2 uses the input variables such as HBsAg, Anti-HBs or HBsAb, Anti-HBc or HBcAb, HBV DNA and Anti-HBcAg-IgM. The layer 1’s output is also used as the input for layer 2. The layer 1’s output is either yes or no as this layer only detects whether the individual suffering from this virus or not. Likewise, layer 2 gave the output as no HBV, acute disease or chronic disease. The accuracy of the system is also evaluated with test cases. Various parameters used to calculate the performance of the system are also determined. The classification accuracy of this system is 94%.