EMG Signal Classification Using Discrete Wavelet Transform and Rotation Forest

Abstract

Electromyographic (EMG) signals are used for the diagnosis of neuromuscular disorders. We have used machine learning algorithms in diagnosing neuromuscular disorders as a decision support system. Hence, in this study, for feature extraction DWT has been used and the Rotation Forest ensemble classifier has been used for classification. Furthermore, we also investigated the performance of different classifiers with Rotation Forest. The performance of a classifier is enhanced using the Rotation Forest ensemble classifier. Significant amount of performance improvement was achieved with a combination of Discrete Wavelet Transform (DWT), and Rotation Forest using k-fold cross validation. Experimental results show the feasibility of Rotation Forest, and we also derive some valuable conclusions on the performance of ensemble learning methods for diagnosis of neuromuscular disorders. Results are promising and showed that the ANN with Random Forest ensemble method achieved an accuracy of 99.13%.