The study of principle component of the surface electromyography signal of the Bicep Brachii muscle

In classification the sEMG pattern, the data of interest can be viewed as drawn from a mixture of geometry or statistical model instead of a single one. Such data are often referred to in different context as `mixed´ or `multi modal´ or `heterogeneous´ or `hybrid´. However, the problem lies on the complexity of the data that represent the behavior of the muscle activity. This paper presents an investigation to determine the principal component of the Bisep Brachii muscle´s EMG features during lifting activity. For this reason principle component analysis is employed to determine dominant, non-dominant and potential EMG features. From the analysis, MaxPS is shown to be the most reliable features with the lowest variance when compared to others. In addition, IEMG, RMS, SSI, WL and MAV also have great potential to represent EMG activity. Meanwhile, results show that MDF and MNF are not useful to represent the EMG activity.