Multi-Dimension Incident Recommendation for Network Maintenance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network operations face low accuracy in recommending similar incidents due to single-dimension comparison, leading to ineffective reference solutions for maintenance personnel.
Innovation Solution
A method for recommending similar incidents through multi-dimension comparison, involving obtaining alarm information, extracting M-dimensional incident diagnosis information, calculating similarity degrees, and filtering historical incidents to provide accurate references for network maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If single-dimension comparison is used to recommend similar incidents, then the system complexity is low, but the recommendation accuracy is low
Solution Approach 1:
The patent transitions from single-dimension comparison (alarm proportion only) to multi-dimension comparison by introducing additional dimensions such as alarm type, affected service, network element, and time characteristics. This dimensional expansion enables more accurate incident similarity assessment while maintaining manageable system complexity through structured comparison frameworks.
Solution Approach 2:
The patent segments the incident comparison process into distinct dimensional components (alarm dimension, service dimension, network element dimension, time dimension). Each dimension is evaluated separately and then integrated to form a comprehensive similarity assessment, allowing the system to achieve high accuracy without overwhelming complexity.
2Measurement precision
If multi-dimension comparison is used to recommend similar incidents, then the recommendation accuracy is improved, but the calculation complexity increases
Solution Approach 1:
The patent divides the complex multi-dimension comparison into separable dimensional evaluations. Each dimension (alarm, service, network element, time) is processed independently using standardized comparison rules, then results are aggregated. This segmentation reduces calculation complexity by avoiding monolithic multi-parameter optimization.
Solution Approach 2:
The patent transforms qualitative incident characteristics into quantifiable parameters suitable for computational comparison. By defining specific measurement parameters for each dimension (e.g., alarm type matching, service impact correlation, time window overlap), the system enables efficient calculation while maintaining high measurement precision.
3Productivity
If monitoring personnel process each incident independently without knowledge accumulation, then the system is simple to operate, but the network maintenance efficiency is low
Solution Approach 1:
The patent implements feedback mechanisms where historical incident data and resolution outcomes are continuously accumulated and fed back into the recommendation system. This creates a learning loop that improves recommendation accuracy over time, enhancing maintenance efficiency while the system autonomously manages the increasing complexity of knowledge accumulation.
Solution Approach 2:
The patent enables the system to automatically accumulate and organize incident knowledge without requiring manual intervention for each new incident. The system self-services by automatically comparing new incidents against historical data, generating recommendations, and updating its knowledge base, thereby improving efficiency while keeping operational complexity low for human personnel.
Data Source
AI summary
A system and method for recommending a similar incident in operations technologies and a related device are provided. The method includes: obtaining alarm information of a to-be-processed incident; obtaining incident diagnosis information of M dimensions based on the alarm information, wherein the M dimensions include M different perspectives of incident diagnosis, and M is an integer greater than 1; performing processing based on the incident diagnosis information of the M dimensions, to obtain a feature of the to-be-processed incident; obtaining a plurality of similarity degrees through calculation based on the feature of the to-be-processed incident and features of a plurality of historical incidents, wherein the plurality of similarity degrees represent respective similarity degrees between the to-be-processed incident and the plurality of historical incidents; and obtaining a similar incident from the plurality of historical incidents through filtering based on the plurality of similarity degrees, and recommending the similar incident.