A method and device for mining frequent symmetric patterns in time series

Abstract

The embodiment of the invention provides a time sequence frequent symmetric pattern mining method and device, and the method comprises the steps: carrying out the analysis of a time sequence accordingto a dynamic programming algorithm and a quasi-Newton method, and determining the key point information of the time sequence; dividing the time sequence into a plurality of time sub-sequences according to the time sequence key point information and a preset interception interval, and performing symmetry analysis on each time sub-sequence to obtain a symmetrical sub-sequence set; analyzing the similarity of each symmetrical sub-sequence in the symmetrical sub-sequence set to obtain a similarity matrix of each symmetrical sub-sequence, and performing density clustering according to the similarity matrix to obtain frequent symmetrical mode information. After key points of the time sequence are determined through a dynamic planning key point detection algorithm, an original time sequence is divided into the time sub-sequences with different lengths, clustering is performed according to similarity, so that a frequent symmetric mode of the time sequence is obtained, and the mining accuracyand diversity are effectively improved.

Description

technical field [0001] The invention relates to the technical field of information processing, in particular to a method and device for mining frequent symmetric patterns of time series. Background technique [0002] Since the beginning of the 21st century, driven by the rapid development of computer science and technology, a large amount of data has been derived in many fields such as society, science, economy, and industry. Analyzing and mining the laws contained in these data is of great significance for guiding production. Time-series data is a sequence of data in which different values ​​of a certain state or variable at different times are arranged in chronological order. In data mining, time series mining is a very important research module, and its application value and research value are very extensive in actual production. [0003] Now time series pattern mining has become an important research direction. Among many patterns, symmetric patterns are especially wide...

AI Technical Summary

Problems solved by technology

However, there may be many sub-patterns in a long-term series, and their symmetry and cycle length are different from each other, so the unified analysis effect is poor

Images