Classification of EMG signals from forearm muscles as automatic control using Naive Bayes

The wheelchair is still a mobility aids commonly used by patients with muscle weakness or stroke patients. Some stroke patients, having constraints in moving a joystick or controlling an electric wheelchair due to muscle limitations of their hands Myo-armband, as wearable device that have an Electromyogram sensor can be used as an alternative in controlling the electric device like wheelchair more easily. The Electromyography Research (EMG) on feature of particular muscle activation pattern which has correlation with a motion contributes inspiration to be applied as motion control media on electric wheelchair. Classification process of EMG will be a new alternative to control wheelchair movement for user or patient who hasn't latitude to move their limb and just able to do easy motion using their forearm. The stages of this project is detecting signal in the muscle using EMG, extracting feature of muscle response in time domain base, and be classified by Naïve Bayes, the dataset classification is pinned in raspberry and output to arduino controller to be used as output motion in motor of electric wheelchair. The result of this research is classification of MAV feature, Peak number, RMS and Gradient Magnitude in 275 stream of muscle data show that detected and correctly can be discriminate 90.18%, thus, a sum of 248 instances and wrongly 9.8182% a sum of 27 instances.