Time series mining based on piecewise aggregate time warping distance
Libin Yang · Journal of Shandong University · 2011
To measure the similarity of time series,a method of time series mining based on piecewise aggregate time warping distance was proposed.Piecewise aggregate approximation was used to transform the data and to extract features from time series so as to reduce dimensionality.After completing the transformation of time series,dynamic time warping was applied to measure the distance between two time series.The proposed method is easy to carry out and the classification result demonstrates that it gets approximately 50% decline of classification error rate of the traditional segmented time warping distance.The new method also has a good performance at running time and clustering in data mining.