A Novel Single-Input Rule Module Connected Fuzzy Logic System and Its Applications to Medical Diagnosis

Abstract

The fuzzy logic system is an expert system based on fuzzy logic and fuzzy reasoning. It consists of a series of fuzzy rules, which can better represent expert knowledge and effectively treat uncertainty and inaccuracy existing in the field of medical diagnosis. To overcome the rules explosion problem, this paper proposes a novel single-input rule module connected fuzzy logic system, in which the rules adopt single-input antecedents and multi-output consequents with the form of Fuzzy Weighted Linear Function (SIRM-FWLF for short). The updating formulas using the steepest descent algorithm are also deduced to optimize the system parameters. At last, numerical simulations are conducted in the MATLAB environment to compare our SIRM-FWLF model with Mamdani, TSK fuzzy logic systems and fuzzy functional SIRM inference model for medical diagnosis of breast cancer. The simulation results show that the proposed SIRM-FWLF FLS can provide higher diagnostic accuracy and recall rates than other fuzzy logic systems or fuzzy inference models.