Combining multiple support vector machines for boosting the classification accuracy of uterine EMG signals

A defining feature of physiological systems is the complexity both in their structures and functions. As a result, classifying physiological data is a difficult task. In this paper, we propose the use of a committee machines with a Support Vector Machines (SVM) as the component classifier in order to boost the classification accuracy of multichannel uterine electromyogram (EMG) signals. The approach was applied on each channel and a majority voting rule was used in order to determine the final decision of the committee. The results indicate that a committee machines exhibits performance unobtainable by an individual committee member on its own. We conclude that this approach can improve the recognition accuracy and has a competitive and promising performance.