Monitoring System Detection Possibility Judgment for Confidence Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monitoring systems fail to detect certain events due to undetectable condition events, leading to lower confidence calculations and potential oversight of root cause analysis, especially when nodes are in shutdown or not generating expected events.
Innovation Solution
The monitoring system includes a detection possibility judgment mechanism that assesses whether a condition event can be detected based on event occurrence rules and performance information, adjusting confidence calculations by considering undetected events and their detection possibilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the monitoring system calculates degree of confidence based on detected condition events only, then the calculation is simple, but the accuracy decreases when events are undetected due to node shutdown or other conditions
Solution Approach 1:
The system performs preliminary detection possibility judgment for each condition event before calculating the degree of confidence. This preliminary assessment determines whether an undetected event should be considered as detected based on predefined rules (e.g., whether the node apparatus is in shutdown state). By performing this action in advance, the system avoids the need for complex post-calculation adjustments and maintains accurate confidence scores without excessive complexity.
Solution Approach 2:
The invention introduces an intermediary mechanism - the detection possibility judgment - that acts as a mediator between event detection and confidence calculation. This intermediary layer evaluates whether undetected events should be considered detected based on contextual information (node status, event type), thereby bridging the gap between simple detection and accurate confidence measurement without requiring complete system redesign.
2Measurement precision
If the monitoring system considers all condition events including undetected ones, then the confidence calculation becomes more accurate, but the computational load increases
Solution Approach 1:
The system performs preliminary detection possibility judgment to pre-classify condition events into detected, undetected-but-considered-detected, and undetected-considered-undetected categories. This preliminary classification avoids the need for exhaustive computational analysis of all possible events, reducing computational load while maintaining accurate confidence calculation by only considering relevant events.
Solution Approach 2:
Instead of performing complete analysis on all condition events, the system applies partial action by selectively considering only those undetected events for which detection possibility judgment returns positive. This partial consideration approach balances computational efficiency with accuracy, avoiding excessive computation on events that would not affect the confidence calculation anyway.
3Reliability
If the monitoring system uses strict event detection rules, then false positives are reduced, but events during node shutdown are incorrectly excluded from analysis
Solution Approach 1:
The system dynamically adjusts event detection behavior based on node operational states. The detection possibility judgment mechanism evaluates contextual information such as whether the node apparatus is in shutdown state and modifies the detection outcome accordingly. This dynamic approach allows the system to maintain strict detection rules for active nodes while adaptively considering events from shutdown nodes, thereby resolving the contradiction between reliability and adaptability.
Solution Approach 2:
The invention changes the detection parameter from binary (detected/not detected) to ternary (detected, undetected-but-considered-detected, undetected-considered-undetected) based on node operational state. By introducing this parameter change, the system can differentiate between events that should be strictly detected and events that should be considered detected despite not being physically detected, thereby adapting to different node states while maintaining detection reliability.
Data Source
AI summary
A monitoring system is configured to perform a cause analysis of an event which occurs in any of a plurality of node apparatuses. The monitoring system is configured to store a plurality of rules indicating a correspondence relationship between one or more condition events regarding any of the plurality of node apparatuses, and a conclusion. The monitoring system is configured to specify the first conclusion associated with the detected condition event based on the rule, and perform a detection possibility judgment on whether the monitoring system can detect the condition event or not, for each of one or more condition events out of a plurality of condition events, and calculate the first index value indicating the certainty of the first conclusion being the cause, based on the existence of detection of one or more condition events associated with the first conclusion and the result of the detection possibility judgment.