Data anomaly detection method and power grid data anomaly detection method

Abstract

The invention discloses a data anomaly detection method. The method comprises the steps of generating an initial model; using a multi-scale generator; training the initial model; generating a loss function; and performing anomaly detection. According to the method, data information of a time sequence is extracted through the sliding window, and hole convolution is used, so that the accuracy of the model is improved, and the generalization ability of the model is improved. The invention further provides a power grid data anomaly detection method. According to the invention, a plurality of generators and a single discriminator are utilized to alleviate the problem of mode collapse. Each generator comprises a convolutional neural network of a different size so as to acquire fine-grained and coarse-grained information of the time sequence. The generator comprises a Transform module which is used for processing the time sequence data so as to improve the precision; according to the method, the generators are balanced by using an attention mechanism, so that currently used data can be well adapted. Therefore, the problems of low precision, poor generalization ability and the like in streaming data anomaly detection can be effectively solved.

Description

technical field [0001] The invention belongs to the field of big data processing, and in particular relates to a data anomaly detection method and a power grid data anomaly detection method. Background technique [0002] Anomaly detection through big data can save manpower and material resources, so it is now more and more common to use data anomaly detection methods to solve problems. Data anomaly detection methods generally use time series anomaly detection. Time series anomalies are observations that are particularly different from other observations in a particular time series. Anomaly detection methods have played an important role in many fields such as extreme weather or climate detection, network intrusion detection, chemical fault diagnosis and power grid fault diagnosis. For example, in extreme weather, quantitative indicators such as wind direction, wind speed, and precipitation will have different degrees of abnormality. Anomaly detection uses models to predict...

AI Technical Summary

Problems solved by technology

Therefore, in order to obtain good experimental results, it is necessary to find a model that can perfectly fit the real data, but when the situation is complex and the data is affected by many factors, it is difficult to use a single model to describe the data in the real world

Images