A Pattern Distance-Based Evolutionary Approach to Time Series Segmentation

Time series segmentation is a fundamental component in the process of analyzing and mining time series data. Given a set of pattern templates, evolutionary computation is an appropriate tool to segment time series flexibly and effectively. In this paper, we propose a new distance measure based on pattern distance for fitness evaluation. Time sequence is represented by a series of perceptually important points and converted into piecewise trend sequence. Pattern distance measures the trend similarity of two sequences. Moreover, experiments are conducted to compare the performance of pattern-distance based method with the original one. Results show that pattern distance measure outperforms the original one in correct match, accurate segmentation.