Extrapolating Associations for Behavior-Deficient Items
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data mining methods face challenges in creating reliable associations for new or unpopular items due to insufficient behavioral data, leading to the 'cold-start' problem, where new items remain unnoticed due to lack of exposure, and content-based associations are less reliable and less predictive of user preferences.
Innovation Solution
The system extrapolates behavior-based associations to 'behavior-deficient' items by using content-based associations, creating new associations based on substitutability between items, allowing new or unpopular items to inherit associations from similar items with established behavioral data, thereby enhancing their exposure and relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If behavior-based associations are used to recommend items, then recommendation accuracy is improved, but new or unpopular items cannot be recommended due to insufficient behavioral data
Solution Approach 1:
The patent introduces content-based associations as an intermediary mechanism to bridge the gap between user behavior data and new items. By using item content attributes (descriptions, categories, tags) as a mediator, the system can generate recommendations for behavior-deficient items without directly relying on their behavioral data, thus resolving the contradiction between recommendation accuracy and new item coverage
Solution Approach 2:
The system performs preliminary action by pre-computing content-based associations for all items before they accumulate behavioral data. This allows new items to immediately inherit associations from similar items based on content similarity, enabling them to be recommended from the moment they are added to the catalog, rather than waiting for behavioral data to accumulate
2Adaptability or versatility
If content-based associations are used to recommend new items, then coverage of new items is improved, but recommendation reliability deteriorates due to lower predictive power
Solution Approach 1:
The patent merges behavior-based associations and content-based associations into a unified recommendation framework. By combining these two association types with different weights, the system leverages the high reliability of behavior-based associations for popular items while utilizing content-based associations to extend coverage to new items, achieving both accuracy and versatility simultaneously
Solution Approach 2:
The system applies local quality by using different association types for different item populations. Behavior-based associations are used for popular items where behavioral data is abundant, while content-based associations are used for new or unpopular items where behavioral data is scarce. This localized application of different association strategies optimizes recommendation quality for each item category
3Reliability
If behavioral data is collected for all items, then association reliability is improved, but data collection time increases for new items
Solution Approach 1:
The system performs preliminary action by pre-computing content-based associations when items are first added to the catalog, eliminating the need to wait for behavioral data to accumulate. This allows new items to immediately participate in recommendation systems with pre-established associations based on their content attributes, significantly reducing the time required before they can be effectively recommended
Data Source
AI summary
Behavior-based associations, such as item-to-item or query-to-item associations, are extrapolated to other items to create new associations. The items to which the associations are extrapolated may be “behavior deficient” items, or items for which the quantity of collected user activity data is insufficient to create meaningful or reliable behavior-based associations. The behavior-based associations are extrapolated based on content-based associations, or another type of “substitutability” association, between items. The items can be any type of item (e.g., products, web sites, documents, etc.) for which user behaviors (e.g., purchases, accesses, downloads, etc.) can be monitored and analyzed to detect behavior-based associations, and for which item content or other available information can be used to assess item substitutability.