A Time Series Similarity Measurement Method Based on Adaptive Segmented Statistical Approximation

Abstract

The invention discloses a time series similarity measurement method based on adaptive segmental statistical approximation. First, the turning point is identified based on the time series code, and the time series is divided into subsequences containing complete fluctuation trends; and then each subsequence is sequentially extracted A variety of statistical features of the local pattern feature vector is constructed; finally, the normalized distance is used to calculate the distance between the local pattern feature vectors to achieve local pattern matching, and use this as a subroutine of the dynamic programming algorithm to achieve global pattern matching. The present invention is superior to other measurement methods in terms of measurement accuracy and calculation efficiency to a greater extent, and can play an important role in people's daily activities and industrial production, such as in financial transactions, traffic supervision, air quality and temperature monitoring, industrial In applications such as process monitoring and medical diagnosis, similarity query, classification, clustering, prediction, anomaly detection, and online pattern recognition are performed on large-scale sampled data or high-speed dynamic data streams.

Description

technical field [0001] The present invention relates to the fields of database, data mining, machine learning, information retrieval, etc., and especially relates to time series data analysis and mining. Background technique [0002] Time series widely exist in people's daily life and industrial production, such as real-time transaction data of funds or stocks, daily sales data in the retail market, sensor monitoring data in the process industry, astronomical observation data, aerospace radar, satellite monitoring data, real-time Weather temperature and air quality index, etc. The industry has proposed many time series analysis methods so far, including similarity query methods, classification methods, clustering methods, forecasting methods, anomaly detection methods, etc. Among them, many methods need to judge the similarity of time series, such as kNN classifier, k-means clustering method, etc. Therefore, the time series similarity measurement method has a wide range of ...

AI Technical Summary

Problems solved by technology

Such methods include segmented aggregation approximation, segmented linear approximation, symbolic aggregation approximation, singular value decomposition, principal component analysis, etc. The first three methods need to segment the original time series first, and then process each sub-segment separately : The segmented aggregation approximation is to calculate the average value of each segment; the segmented linear approximation is to do line segment fitting on each segment; The extracted features are relatively single, which makes it less capable of expressing time series fluctuation patterns

Singular value decomposition and principal component analysis are realized by performing a unified eigenmatrix decomposition on all time series; the typical defects of these two types of methods are that they have high computational complexity, and the decomposition process can only be done in memory, and the data Very low scalability at scale

Images