Abnormal point detection model verification method and device, computer equipment and storage medium

Abstract

The invention discloses an abnormal point detection model verification method and device, computer equipment and a storage medium. The method comprises the following steps: constructing an isolated forest model according to a preset current abnormal point proportion and a to-be-classified sample; Classifying the samples to be classified according to the isolated forest model and the current abnormal point proportion to obtain a classification result; Classifying the classification result according to a preset data attribute classification strategy to obtain an attribute classification result;Obtaining each attribute classification result in the attribute classification results; And if the first attribute average value in each attribute classification result is smaller than the second attribute average value, adding a reasonable model label to the isolated forest model. According to the method, an intelligent decision is adopted to automatically, quickly and accurately judge the classification reasonability of the isolated forest model, and low efficiency and high cost caused by manual judgment are avoided.

Description

technical field [0001] The invention relates to the technical field of intelligent decision-making, in particular to a verification method, device, computer equipment and storage medium for an abnormal point detection model. Background technique [0002] Outlier analysis is the process of checking whether the data has input errors or contains unreasonable data. It is very disadvantageous to ignore the existence of outliers. Including outliers in the calculation and analysis process of data without eliminating them will have adverse effects on the results. Influence. [0003] The unsupervised anomaly detection model is commonly used in outlier analysis, and for the output results of the unsupervised anomaly detection model, it is often necessary to judge the rationality of the model classification from reality. The most commonly used method is the expert method, that is, to consult relevant experts to judge whether the model classification is reasonable based on their busine...

AI Technical Summary

Problems solved by technology

[0004] The embodiment of the present invention provides a verification method, device, computer equipment and storage medium for an anomaly detection model, aiming at solving the problem of unsupervised anomaly detection model output results in the prior art through relevant experts using their business knowledge to judge the model Whether the classification is reasonable, which leads to the problem of high judgment cost and low efficiency

Images