An efficient Parkinson disease diagnosis system based on Least Squares Twin Support Vector Machine and Particle Swarm Optimization

This paper presents an efficient Parkinson disease diagnosis system using Least Squares Twin Support Vector Machine (LSTSVM) and Particle Swarm Optimization (PSO). LSTSVM is a promising binary classifier and has shown better generalization ability and faster computational speed. PSO is used for feature selection and parameter optimization. Parkinson disease dataset is taken from UCI repository. The performance of proposed system is compared with other existing approaches in terms of accuracy, sensitivity and specificity. Experimental results validate the effectiveness of proposed Parkinson disease diagnosis system over other exiting techniques.