Multi-classification techniques applied to EMG signal decomposition

In this paper we study the effectiveness of using multiple classifier combination for EMG signal decomposition aiming to obtain more accurate results than is possible from each of the constituent classifiers. The developed system employs an ensemble of error-independent modified certainty classifiers fused at the abstract and measurement levels for integrating information to reach a collective decision. For decision combination at the abstract level, the majority voting scheme has been investigated. While at the measurement level, two types of combination methods have been investigated: one used fixed combination tides that do not require prior training and a trainable combination method. For the second type, the fuzzy integral method was used. The ensemble classification task is completed by feeding the classifiers with different features extracted from the EMG signal. The results show that using classifier fusion methods improved the overall classification performance.