Algorithm for Finding Similar Patterns over Time-Series Data Based on Wavelets and Feedback

To obtain valuable similar time-series, the advantage of the wavelet transform and its performance in finding similar patterns over time-series data are analyzed. The disadvantages of existing algorithms based on wavelet transform are summarized. An algorithm for finding similar patterns over time-series data based on wavelets and feedback is put forward, besides the method for validating results based on the depth of verification. The validity and practicability of the algorithm are proved through experiments.