Classification of electromyogram using weight visibility algorithm with multilayer perceptron neural network

Classifing of electromyographic (EMG) signal has been a significant issue on diagnosis for the disease since the signal is complex and non-stationary. The key on the classification is feature extraction. In this paper we propose a novel feature extraction technique based on transforming the signal to complex network via weight visibility algorithm. The feature vector is obtained from statistical mechanics of complex network. Then, multilayer perceptron neural network is employed for classification. The proposed method classified the signals into 3 cases, i.e., healthy, myopathy, and neuropathy. The experimental results show that the proposed method identified and classified the EMG signal with average accuracy of 94.75%.