Classification of Voice Signals through Mining Unique Episodes in Temporal Information Systems: A Rough Set Approach.
Krzysztof Pancerz, Wiesław Paja, Mariusz Wrzesień, Jan Warchoł · 2012
Abstract. Classification of voice signals in a time domain for detecting some disturbances can be made through mining unique episodes in temporal information systems. In ideal case, a voice signal is periodic and signal shapes in each period are the same. In case of larynx diseases, some disturbances in a voice signal can be distinguished. Selected time windows (sequences of signal samples) of a voice signal constitute a temporal information system. Appearing unique episodes in such a system can indicate disturbances in the signal. In the paper, we show the methodology, based on a rough set approach, of mining unique episodes in temporal information systems as well as its application for classification of voice signals.