Unlock instant, AI-driven research and patent intelligence for your innovation.

Log anomaly detection method combining bidirectional sliced ​​GRU and gated attention mechanism

An anomaly detection and attention technology, applied in neural learning methods, error detection/correction, hardware monitoring, etc., can solve problems such as large time overhead and performance degradation, and achieve high accuracy, simple parameters, and complex parameter resolution

Active Publication Date: 2022-04-19
OCEAN UNIV OF CHINA +1
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the huge amount of log data, performance degradation and huge time overhead

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Log anomaly detection method combining bidirectional sliced ​​GRU and gated attention mechanism
  • Log anomaly detection method combining bidirectional sliced ​​GRU and gated attention mechanism
  • Log anomaly detection method combining bidirectional sliced ​​GRU and gated attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0077] Such as Figure 1 to Figure 8 Shown is an embodiment of the log anomaly detection method combining bidirectional slice GRU and gated attention mechanism of the present invention, including the following steps:

[0078] S101, using Spell to parse out the logkey from the log data, and using the Word2Vec tool to train the log key vector;

[0079] S102, converting the logkey into a fixed-length index, each index corresponding to a logkey sequence vector; splicing the logkey sequence vector into a logkey sequence matrix, as the embedding layer weight of the model;

[0080] S103. Slicing the logkey obtained from log parsing as the input of Bi-SSGRU-GA-Attention model;

[0081] S104, input the log key minimum subsequence index representation into the embedding layer, and then input it into the Bi-SSGRU layer to extract the logkey subsequence hierarchical features;

[0082] S105, inputting the features extracted by each subsequence through the Bi-SSGRU to the GA-Attention lay...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention belongs to the technical field of log anomaly detection, and discloses a log anomaly detection method combining a bidirectional slicing GRU and a gated attention mechanism. , introduce bidirectional slicing and gated attention mechanism to build a log anomaly detection model, and use the parsed feature sequence as the input of the log anomaly detection model to train the log anomaly detection model, and use the trained log anomaly detection model for log anomaly detection. . The log anomaly detection algorithm of the present invention has the advantages of simple parameters and fast convergence speed, and at the same time reduces the running time, achieves a high accuracy rate, and achieves a relatively ideal effect in log analysis of large-scale information systems.

Description

technical field [0001] The invention belongs to the technical field of log anomaly detection, and in particular relates to a log anomaly detection method combining a bidirectional slice GRU and a gated attention mechanism. Background technique [0002] At present: Various logs are generated during system operation, which record the state of the system during operation and various operations performed by the system, and are a good source of information for online monitoring and anomaly detection. Therefore, it is of great significance to maintain the security and stability of the system to quickly and accurately detect the abnormal logs existing in the system. [0003] System log anomaly detection has been a hot research topic in the field of anomaly detection. System logs are composed of a variety of unstructured data sets in non-fixed formats, and are closely related to many disciplines such as statistics, natural language processing, and machine learning. In recent years...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F11/34G06K9/62G06N3/04G06N3/08
CPCG06F11/3476G06N3/08G06N3/045G06F18/2411
Inventor 顾士景马超张闻彬王高洲殷齐林郭爽爽黄振刘荫韩圣亚汤琳琳于航徐浩张悦王惠剑郑海杰张凯刘培顺
Owner OCEAN UNIV OF CHINA