Robust Feature Combination for KNN Classification to Obtain Classification Accuracy

A popular method for acquiring the electrical activity of muscle is called Electromyography (EMG). It is widely used for the detection and treatment of neuromuscular disorders. EMG has been in use in medical field for the diagnosis of neuromuscular disease for several years. Neuromuscular disorder that affects the peripheral nervous system includes Myopathy and Amyotrophic Lateral Sclerosis (ALS). Primary goal of this paper is to classify normal, ALS and myopathy EMG signals based on the features being extracted. Different combinations of frequency domain, time domain and time- frequency domain features were obtained and analyzed for the best result. About 97% classification accuracy was achieved when the feature set comprising of standard deviation and Willison amplitude was classified using K-Nearest Neighbor (k- NN) classifier.