A Design of Fuzzy Classifier Using Fuzzy Clustering and Voice Analysis for Diagnosis of Parkinson Disease

Abstract

This paper is concerned with a design of fuzzy clustering based fuzzy classifier based on voice database of Parkinsons patients and healthy people for diagnosis of Parkinsons Disease (PD). For this purpose, we use a database of UCI (University of California Irvine) machine learning repository for voice disorder detection of Parkinsons patients and design automatic fuzzy if-then rules and classifier based on fuzzy clustering. This database consists of 23 Parkinsons patients and 8 healthy persons, including 16 input variables with voice features and two classes. The experimental results on the database revealed that the proposed fuzzy classifier showed good classification performance in comparison to pervious RBFC (Radial Basis Function Classifier), NFC(Neuro-Fuzzy Classifier), and ANFC (Adaptive Neuro-Fuzzy Classifier).