Data abnormal mutation point detection algorithm based on limited accompanying group mechanism

Abstract

The invention discloses a novel abnormal mutation point detection method, and particularly relates to a data abnormal mutation point detection algorithm based on a limited accompanying group mechanism. A finite accompanying group mechanism is utilized to help to efficiently and accurately identify and detect the mutation points in the industrial data stream. According to the method, a limited accompanying group thought is provided, and effective identification and detection of abnormal mutation points of the data stream are realized by selecting a limited high-quality accompanying group and constructing a to-be-detected point group. The algorithm has low delay, and abnormal mutation points can be timely identified in real-time data streams.

Description

technical field [0001] The invention relates to a novel data abnormal mutation point detection algorithm, in particular to a data abnormal mutation point detection algorithm based on a limited jury mechanism. Background technique [0002] Since the concept of the Industrial Internet was proposed, the industry has stored a large amount of data related to equipment runtime through data collection and big data transmission and storage technologies. In terms of ensuring the safe operation of industrial equipment, in addition to manual maintenance and repair of equipment, the industry has been constantly trying to use existing historical data and real-time monitoring data streams of equipment to conduct related research to achieve more efficient equipment monitoring and data anomalies diagnostic function. At the same time, more and more scholars have begun to study fault diagnosis algorithms based on industrial data in order to provide a more efficient and accurate diagnosis alg...

AI Technical Summary

Problems solved by technology

However, traditional detection methods are difficult to meet the needs of real-time industrial data flow anomaly detection with large amounts of data.

For example, statistical methods are suitable for detecting discrete and sudden value anomalies in the introduction, but it is difficult to effectively identify continuous abnormal sequence intervals

Clustering-based methods quantify the distance between abnormal points and normal point clusters to judge outliers. The computational complexity of different clustering models is different, and the detection results are more dependent on the quality of clustering. At the same time, it is not suitable for large-scale Anomaly detection on datasets and real-time data streams

The method based on similarity measurement calculates the similarity between standardized sequences to judge whether there is abnormal data, but this method has a large time cost and is not time-sensitive

In the method based on rule constraints, the researchers proposed that order dependence and speed constraints can effectively use the timing characteristics in time series to repair highly abnormal data, but this method is usually difficult to meet the needs of sequence anomaly detection with variable patterns

Images