Disk anomaly detection method based on neighborhood partitioning and isolation reconstruction

Abstract

The embodiment of the invention provides a disk anomaly detection method based on neighborhood partitioning and isolation reconstruction, which comprises the following steps: collecting disk SMART information, screening effective disk feature attributes to form a data set, and carrying out exponential smoothing processing on the data set to obtain a disk training set; randomly sampling the training set for multiple times to obtain a plurality of sub-training sets, constructing a disk feature isolation region in the sub-sets by taking the distance from each point to the nearest point as a radius, and taking test points which do not belong to any region as global anomalies; for a non-global abnormal test point, using the radius ratio of the area where two continuous neighbor points are located as a front abnormal value of the test point in the area; reconstructing an area after a test point is included, and taking a radius ratio of the area where the test point is located before and after reconstruction as a post abnormal value of the test point in the area; and obtaining an exception score by combining previous and later exception values of all areas where test points are located. According to the technical scheme provided by the embodiment of the invention, the abnormal disk recall rate can be effectively improved.

Description

【Technical field】 [0001] The invention relates to an anomaly detection method in the field of machine learning, in particular to a disk anomaly detection method based on neighborhood partitioning and isolation reconstruction. 【Background technique】 [0002] At present, disks are the most widely used computer storage data, and the operation of disks is directly related to the security of stored data. Data centers generally have hundreds or thousands of disks, which greatly increases the possibility of system failure. Therefore, the data center needs to adopt some mechanisms to detect the abnormality of the disk, so as to avoid irreversible damage or loss of data. [0003] Currently, the commonly used disk anomaly detection method is the threshold detection method based on SMART data. It can monitor and record the hardware operation of the disk itself by sending detection instructions in the disk, and compare it with the preset safety value set by the manufacturer. If it is...

AI Technical Summary

Problems solved by technology

[0006] In view of this, the embodiment of the present invention proposes a disk anomaly detection method based on neighborhood partitioning and isola

Images