Time-Series Event Classification With Selective Manual Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing analysis systems for plant management face a significant workload burden due to the complexity of associating data and events, leading to increased management costs and inefficiencies.
Innovation Solution
An analysis system that includes a classifier for event type classification of time-series data, a display information generation unit for visualizing undecided event types, and an input unit for associating event types, allowing for reduced management burdens through improved data association and event determination processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual association of data and events is performed, then accuracy of event classification can be ensured, but management burden and workload increase significantly
Solution Approach 1:
The system segments the event classification task into two parts: automated classification by the analysis unit for initial categorization, and manual verification only for undecided events. This segmentation reduces the management burden by automating routine classifications while maintaining accuracy through selective manual review of ambiguous cases.
Solution Approach 2:
The analysis unit acts as an intermediary between raw time-series data and manual event association. It performs preliminary classification and presents only undecided events to the user, reducing the workload while maintaining classification accuracy through this intermediate processing step.
2Device complexity
If automated classification is applied to all time-series data, then management burden is reduced, but classification accuracy may deteriorate due to undecided events
Solution Approach 1:
The system segments time-series data into two categories: decided events (automatically classified) and undecided events (requiring manual input). This allows automated processing to handle clear cases, reducing management burden, while preserving the option for manual verification on ambiguous cases to maintain accuracy.
Solution Approach 2:
Instead of requiring full manual verification of all events, the system applies partial automation to decided events and reserves manual action only for undecided events. This partial action approach reduces management burden while maintaining sufficient classification accuracy for the majority of clear-cut cases.
3Measurement precision
If complete manual verification of event types is performed, then classification accuracy is maintained, but time consumption and productivity decrease
Solution Approach 1:
The verification process is segmented into automated classification for decided events and manual verification only for undecided events. This segmentation maintains classification accuracy for clear cases through automation while improving productivity by limiting manual verification to only the necessary ambiguous cases.
Solution Approach 2:
The system extracts and identifies undecided events from the complete dataset, separating them from decided events. This extraction allows automated processing to handle the majority of clear cases efficiently, improving productivity, while maintaining accuracy through focused manual verification on the extracted undecided portion.
Data Source
AI summary
Provided is an analysis system including: an analysis unit including a classifier that performs classification of an event type on input time-series data; a display information generation unit that generates first display information used for displaying, out of the time-series data, first time-series data in which association of an event type is undecided and which is classified by the classifier as a first event type corresponding to a state where a target event is occurring, second time-series data associated with the first event type, and third time-series data associated with a second event type corresponding to a state where the target event is not occurring; and an input unit that accepts first input regarding association of an event type with the first time-series data.