Feature selection of EMG signals based on the separability matrix and rough set theory

Recognizing bio-signals, such as EMG, EEG, EOG and ECG, is a promising theme of study since it provides with a convenient means for human-machine interaction. In the earlier works were proposed various approaches of determining features of bio-signals that are capable of discerning predefined motions/intentions of human, but most of them were only applicable to a single subject due to inherent characteristics of bio-signals. Lately, several structures of pattern classifier with the known features have been proposed to cope with the subject-dependency, but their error rates are still conspicuous in accommodating multiple subjects. Based on the separability matrix and rough set theory, this paper presents a comparative experimental study to minimize the subject-dependency. It is shown that the induced feature set obtained by the proposed feature selection algorithm, has less subject-dependency than other existing methods.