Attribute-Based Association Mining for Catalog Item Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting associations between items, such as in electronic catalogs, face challenges with 'cold-start' issues where new or rarely purchased items lack sufficient behavioral data, leading to low-quality or nonexistent associations, and popular items are often associated with unrelated items due to high usage volume.
Innovation Solution
An attribute-based association mining system that analyzes item attributes, using techniques like latent semantic analysis and singular value decomposition, to detect associations between items by calculating degrees of fit and correlation, even for behavior-deficient items, and distinguishes between contextually similar and duplicate items.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If behavior-based association detection methods are used, then associations can be detected for items with sufficient behavioral data, but new or rarely purchased items lack sufficient behavioral data leading to low-quality or nonexistent associations
Solution Approach 1:
The patent introduces attribute data as an intermediary element to bridge the gap when behavioral data is insufficient. By using item attributes (such as product categories, descriptions, and metadata) as a mediator, the system can generate associations for behavior-deficient items through attribute-based similarity analysis, thereby maintaining association quality even when behavioral data is lacking
Solution Approach 2:
The patent changes the detection parameters by switching from purely behavior-based detection to a hybrid approach that incorporates attribute-based detection. This parameter change enables the system to detect associations for items with insufficient behavioral data by utilizing alternative parameters (item attributes), thus resolving the contradiction between association quality and data availability
2Quantity of substance
If behavior-based association detection is used for popular items, then sufficient behavioral data is available, but popular items are often associated with unrelated items due to high usage volume
Solution Approach 1:
The patent applies local quality by differentiating the association detection approach based on item characteristics. For popular items, it combines behavior-based detection with attribute-based filtering to refine associations locally, ensuring that high usage volume does not lead to spurious associations. The system adjusts the detection strategy according to the specific item's data characteristics
Solution Approach 2:
The patent implements feedback mechanisms where association results are continuously refined. By using attribute-based detection as a feedback layer, the system can validate and correct behavior-based associations, especially for popular items where high traffic might generate noisy data. This feedback loop improves measurement precision by filtering out unrelated associations
3Device complexity
If traditional association detection methods are used, then processing is simpler, but the system cannot distinguish between contextually similar and duplicate items
Solution Approach 1:
The patent segments the association detection process into distinct phases: initial behavior-based detection, attribute-based refinement, and duplicate filtering. This segmentation allows the system to maintain reasonable processing complexity while progressively improving measurement precision at each stage, ultimately enabling distinction between contextually similar and duplicate items
Data Source
AI summary
A method of detecting associations between items can include identifying a plurality of items represented in a data repository from which to select items to recommend to a target user, each item including one or more attributes. A degree of fit between an item's attributes and other items is calculated. The degree of fit can indicate the relevance of the attributes of one item to the other item. A degree of association between the two items is calculated based at least in part on the calculated degree of fit. The degree of association between the two items can indicate the relatedness of the two items. Based on this degree of association, an association between the items can be stored in a data repository.