Error-Contribution Ratio Calculation for Channel Measurement Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional diagnostic models fail to identify the specific equipment causing errors in facilities, leading to increased investigation burdens for field workers as they lack the capability to provide clear clues for error causation.
Innovation Solution
An information processing apparatus and method that calculates an error-contribution ratio for each channel by using channel measurement data and a machine learning model to determine the degree of error contribution, allowing for the identification of the equipment responsible for errors through the determination of a score based on parameter differences and classification boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a conventional diagnostic model is used to determine errors in batches, then error detection capability is provided, but the ability to identify specific error-causing equipment is lost
Solution Approach 1:
The patent segments the overall error detection task into two parts: (1) batch-level error detection using a diagnostic model, and (2) channel-level error contribution analysis using an error-contribution ratio calculation unit. This segmentation allows the system to first identify that an error exists in a batch, then separately determine which specific measurement channels (equipment) contributed to the error, thereby preserving error cause information that would otherwise be lost.
Solution Approach 2:
The patent introduces an intermediary component - the error-contribution ratio calculation unit - that bridges the gap between error detection and error cause identification. This intermediary takes the diagnostic model's error determination as input and produces detailed error contribution information for each measurement channel, acting as a mediator that transforms incomplete error detection results into actionable diagnostic information.
2Measurement precision
If field workers manually investigate error causes without guidance, then thorough investigation can be conducted, but investigation time and burden increase significantly
Solution Approach 1:
The system performs preliminary action by automatically calculating error-contribution ratios for each measurement channel before field workers begin their investigation. This preliminary analysis identifies which channels have the highest error contribution, allowing workers to focus their thorough investigation efforts on the most likely culprits rather than examining all equipment equally, thereby reducing investigation time while maintaining thoroughness.
Solution Approach 2:
The patent implements feedback by providing field workers with quantitative error-contribution ratio data that guides their investigation. This feedback mechanism transforms the investigation process from a blind search to a targeted examination, where workers receive continuous guidance on which channels to prioritize based on the calculated error contributions, reducing both time and burden.
3Device complexity
If no error contribution analysis is provided, then the diagnostic model remains simple, but users lack guidance on where to start investigation
Solution Approach 1:
The patent segments the diagnostic system into a simple error detection component and a separate error analysis component. The diagnostic model itself remains simple and focused on binary error detection, while the error-contribution ratio calculation unit handles the complexity of identifying specific error sources. This segmentation maintains the simplicity of the core diagnostic model while adding investigative guidance capability.
Solution Approach 2:
The error-contribution ratio calculation unit serves as an intermediary that adds investigative guidance without complicating the core diagnostic model. It takes the simple error detection output and transforms it into actionable guidance information, allowing the diagnostic model to remain simple while still providing users with clear direction on where to start their investigation.
Data Source
AI summary
An information processing apparatus includes: a controller that: acquires channel measurement data for each of one or more channels that is a measurement target, and calculates, for each of the one or more channels, an error-contribution ratio based on a score determined for each of parameters extracted from the channel measurement data acquired for each of the one or more channels, the error-contribution ratio indicating a degree by which each of the one or more channels contributes an error, and the score being determined based on a difference between each of the parameters and a classification boundary used by a machine learning model that classifies the parameters into one of an error class and a normal class.