Classification of EMG Signals by K-Nearest Neighbor Algorithm and Support Vector Machine Methods

Abstract

Electromyography (EMG) is a medical measurement system. EMG measurements are required for the diagnosis of some diseases and used in order to facilitate physicians' work. In this study, MUAPs' in an EMG data set that contains both healthy and Amyotrophic Lateral Sclerosis (ALS) disease subjects are represented in time domain and frequency domain with a total of 10 feature vectors. Two pattern recognition methods, namely k-Nearest Neighbor (k-NN) and Support vector machine (SVM) classifier are employed and compared. In terms of classification accuracy, k-NN classifier give slightly higher success rate than SVM classifier for the existing data set and feature vectors.