Promotion Server Multicollinearity Detection for Targeting Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sales promotion systems face difficulties in identifying feature parameters contributing to sales results and determining the effectiveness of promotions when the number of parameters increases, leading to potential multicollinearity issues that complicate analysis and target selection.
Innovation Solution
A promotion support server that acquires and analyzes promotion results data to identify feature parameters in multicollinearity relationships, providing notifications and supporting the selection of target features for promotions, thereby enhancing the precision of sales promotion planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of feature parameters for customer classification is increased to improve promotion targeting precision, then the ability to identify significant features contributing to sales results deteriorates due to multicollinearity
Solution Approach 1:
The patent extracts and identifies multicollinear feature parameters from the dataset using statistical analysis (variance inflation factor calculation). By separating out these problematic correlated features, the system prevents them from being used in promotion targeting, thereby maintaining analysis clarity while preserving the benefits of having many feature parameters for customer classification.
2Adaptability or versatility
If the number of feature parameters is increased to provide more detailed customer classification, then the independence of features deteriorates leading to high correlation between parameters
Solution Approach 1:
The system implements a feedback mechanism that automatically calculates correlation metrics (variance inflation factor) for each feature parameter and provides notifications when multicollinearity is detected. This feedback loop enables the promotion planning system to adaptively identify and exclude correlated features, maintaining feature independence while preserving comprehensive customer classification capabilities.
3Quantity of substance
If multiple correlated feature parameters are used for promotion targeting, then the coverage of customer segments is improved, but the ability to determine promotion effectiveness deteriorates
Solution Approach 1:
The patent extracts multicollinear features through statistical analysis and excludes them from promotion target selection. By removing these correlated parameters that would confound effectiveness measurement, the system maintains comprehensive customer segment coverage through independent features while ensuring clear, unbiased promotion effectiveness evaluation.
Data Source
AI summary
According to an embodiment, a promotion support server includes a storage unit, a communication interface, and a processor. The processor is configured to acquire promotion results data including a plurality of feature parameters for recipients of a promotion related to a product, identify feature parameters in the plurality of feature parameters that are in a multicore relationship, store the identified feature parameters in the storage unit in association with a product identification, cause a display to display a promotion target setting screen on which feature parameters of a recipient of another promotion related to the product can be set by an operator, and cause the display to display a notification when a designated feature parameter among the feature parameters on the promotion target setting screen is one identified as being in a multicore relationship.