Recommendation List Reordering for Item Diversity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing recommendation systems often present users with lists of items that are too similar, limiting their discovery of diverse complementary items, as they primarily rank items based on purchase histories, leading to groups of the same type being displayed together.
Innovation Solution
The system reorders or modifies purchase-history-based related items lists by incorporating item relationships from viewing histories and attribute similarities, ensuring diverse item recommendations by using an overlap score to position dissimilar items adjacent to each other, thereby increasing the variety of items presented.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If recommendation systems rank items based on purchase histories, then the recommendations are relevant to user preferences, but the items displayed are too similar and lack diversity
Solution Approach 1:
The patent applies local quality by differentiating the treatment of items in the recommendation list based on their position. Items at the top of the list are selected for high relevance to user preferences, while items lower in the list are selected to provide diversity. This allows different parts of the recommendation system to serve different functions - the top positions satisfy relevance requirements while lower positions satisfy diversity requirements, thereby resolving the contradiction between relevance and diversity.
2Reliability
If the system presents only highly relevant items, then user satisfaction with relevance is high, but users miss out on discovering diverse complementary items
Solution Approach 1:
The patent segments the recommendation list into different zones or positions, where the top positions are dedicated to highly relevant items that ensure user satisfaction, and lower positions are dedicated to diverse items that enable discovery. This segmentation allows the system to simultaneously provide both high relevance and diversity without compromise, as each segment serves its specific purpose in the overall recommendation strategy.
3Productivity
If the recommendation list is ordered by purchase history frequency, then the most relevant items appear first, but similar items are grouped together reducing variety
Solution Approach 1:
The patent introduces dynamic reordering of the recommendation list based on multiple factors including purchase history frequency and item similarity. Rather than a static ranking, the system dynamically adjusts the order of items to balance relevance and diversity. Items are repositioned in the list based on their similarity to previously displayed items, ensuring that diverse items are strategically placed to maximize variety while maintaining overall relevance. This dynamic approach resolves the contradiction by making the recommendation list adaptable rather than fixed.
Data Source
AI summary
A server system increases the diversity of item recommendations provided to a target user by using item similarity data to reorder a ranked list of recommended items for presentation to a user. The reordering is performed such that items identified as similar to each other are spaced apart from each other by at least a minimum number of positions. This minimum number may be selected based, e.g., on how many recommended items will initially be presented on a requesting user device, which may depend on the display size or other attributes of the user device. The server system generates a user interface that displays an initial portion of the reordered list. The user interface includes controls for scrolling through the reordered list.