Method, device and equipment for detecting anomaly of time series data and medium

Abstract

The invention provides a time sequence data anomaly detection method, device and equipment and a readable medium, and the method comprises the steps: constructing an Autoencoder model, training the Autoencoder model through correct time sequence data, and obtaining a trained Autoencoder model; inputting time sequence data to be detected into the trained Autoencoder model, and obtaining a first abstract feature after the time sequence data is encoded by the Autoencoder model; decoding the first abstract feature through an Autoencoder model to obtain a decoding matrix, inputting the decoding matrix into the trained Autoencoder model, and encoding the decoding matrix through the Autoencoder model to obtain a second abstract feature; and determining whether the to-be-detected time sequence data is abnormal or not based on the relationship between the first abstract feature and the second abstract feature. By using the scheme of the invention, whether the time sequence is abnormal or not can be judged without manually labeling the time sequence data, and whether the time sequence data is abnormal or not can be accurately judged.

Description

technical field [0001] The present invention relates to the field of computers, and more particularly to a method, apparatus, device and readable medium for abnormal detection of time series data. Background technique [0002] Anomaly detection is one of the most common research directions in time series data analysis, which is defined as the process of identifying abnormal events or behaviors from normal time series. Anomaly detection is widely used in many fields of industry, such as quantitative trading, network security detection, self-driving cars and routine maintenance of large industrial equipment. In the case of in-orbit spacecraft, failure to detect hazards could result in serious or even irreparable damage due to their expensive and complex systems. Anomalies can develop into critical failures at any time, so accurate and timely anomaly detection can alert aerospace engineers to take early action. [0003] The problem of anomaly detection in time series is usual...

AI Technical Summary

Problems solved by technology

[0004] Anomaly detection algorithms are divided into two categories. The first category uses classification algorithms to mark each time point as abnormal / non-abnormal, and then classifies each time point through the classification algorithm. The disadvantage is that it needs to analyze the abnormal / non-abnormal Manual labeling has obvious reliance on human judgment; the second type uses prediction algorithms to predict the signal of a certain point, and then tests whether the actual value of the point is different from the predicted value, and then observes whether the difference is enough to treat it as Exception, the disadvantage is that it depends on the accuracy of the budget algorithm

Images