Cause Analysis Using Outlier Scores for Independent Abnormalities
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In a distribution warehouse, identifying the cause of changes to an abnormal state is challenging due to variations in steady states and independent occurrences of abnormalities, unlike manufacturing factories where multiple abnormalities often occur under the same conditions.
Innovation Solution
A cause analyzing apparatus that acquires reference and target data sets including objective and explanatory variables, calculates outlier scores, relationship weights, and property weights, and uses these to determine cause scores for explanatory variables associated with changes in objective variables, facilitating the identification of causes even in independently occurring abnormalities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a statistical test framework calculating p-values from bias in abnormalities is used, then cause identification is effective when multiple abnormalities occur under the same condition, but cause identification becomes difficult when abnormalities occur independently for each condition
Solution Approach 1:
The patent changes the analytical parameters from p-values (statistical significance) to outlier scores (deviation magnitude). This parameter transformation allows the system to effectively measure cause strength whether abnormalities occur multiple times under the same condition or independently, as outlier scores capture the degree of deviation regardless of frequency patterns
Solution Approach 2:
The patent creates a universal cause analysis framework that handles both repeated abnormalities (manufacturing factory pattern) and independent abnormalities (distribution warehouse pattern) through the same outlier score mechanism. The system becomes multi-functional by accommodating different abnormality occurrence patterns without requiring separate analytical approaches
2Productivity
If conventional analysis methods focusing on bias in abnormal numbers are applied, then analysis is effective for manufacturing factories with repeating operations, but analysis effectiveness decreases for distribution warehouses with varying steady states
Solution Approach 1:
Instead of analyzing bias in abnormal numbers (conventional approach), the patent inverts the approach by calculating outlier scores that measure how much target data deviates from reference data distributions. This inversion makes the analysis effective for environments with varying steady states where abnormality patterns are not repetitive
Solution Approach 2:
The patent transforms the analytical parameter from counting abnormal occurrences to measuring distributional deviation through outlier scores. This parameter change enables the system to adapt to different operational environments by focusing on the magnitude of deviation rather than the frequency pattern of abnormalities
Data Source
AI summary
According to one embodiment, a cause analyzing apparatus includes a processing circuit. The processing circuit acquires a reference data group and target data. The processing circuit calculates an outlier score of an explanatory variable of the target data. The processing circuit calculates a relationship weight representing strength of a relationship between an objective variable and the explanatory variable of the reference data group. The processing circuit calculates a property weight representing a degree of match between values of the objective variable and the explanatory variable of the target data. The processing circuit calculates a cause score of the explanatory variable based on the outlier score, the relationship weight, and the property weight.