Time series data feature extraction and classification method

Abstract

The invention discloses a time series data feature extraction and classification method, which comprises the following steps: firstly, carrying out data warping processing on time series data to enable data points to have the same time interval; removing abnormal points existing in the normalized time sequence data; carrying out data supplementation on missing values originally existing in the time sequence data and rejected abnormal values, and carrying out simple moving averaging on the time sequence data; performing stationarity verification on the time sequence data after simple moving averaging so as to judge whether the characteristics of the time sequence data are stationary or fluctuating; segmenting the data according to a set cycle length; performing Pearson's correlation coefficient calculation on every two adjacent time sequence data segments to obtain a Pearson's correlation coefficient array, taking a median in the array as a final correlation coefficient, and when the correlation coefficient is greater than or equal to a threshold value, judging that the time sequence has a corresponding periodic property; and classifying the time series data according to the extracted data features. According to the invention, the accuracy of time series data anomaly detection is improved.

Description

technical field [0001] In particular, the invention relates to a time series data feature extraction and classification method. Background technique [0002] At present, there are many mature anomaly detection algorithms in time series data anomaly detection, but most anomaly detection algorithms are not universal. When the same detection algorithm is faced with different types of time series data, the detection accuracy is often very different. Therefore, it is urgent to propose a method that can extract the characteristics of time series data and classify them, so that the best anomaly detection algorithm can be selected according to different types of time series data, so as to achieve the purpose of improving the accuracy of anomaly detection. Contents of the invention [0003] The purpose of the present invention is to provide a time series data feature extraction and classification method. [0004] The technical scheme adopted in the present invention is: [0005] ...

AI Technical Summary

Problems solved by technology

[0002] At present, there are many mature anomaly detection algorithms in time series data anomaly detection, but most of the anomaly detection algorithms are not universal. When the same detection algorithm is faced with different types of time series data, the detection accuracy is often very different.

Images