Classification of multichannel uterine EMG signals using a reduced number of channels

Abstract

Multisensor recording is an important technique used for solving various pattern recognition problems such as the classification of electrophysiological signals. Studies have shown that, although there is a correlation between the electrical activities recorded at different sites, the characteristics of the recorded signal depend on the position of the recording electrode. In this study, we search for the combination of channels that can provide the highest classification accuracy of multichannel uterine (electromyogram) EMG signals. The procedure of reducing the total number of channels is described. Our approach is tested by calculating 4 statistical measures. Results have shown that our proposed approach is capable of reducing the number of recording channels and improving the global classification accuracy.