Co-Selection Graph for Personalized Content Bundle Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for recommending digital media content and applications often group recommendations by categories or developer, failing to provide a cohesive solution that matches a user's unique interests and changing circumstances.
Innovation Solution
A computer-based system determines co-selection scores for pairs of content items, representing the probability of simultaneous downloads, and outputs a subset of content items connected by edges in a graph with high co-selection scores, recommending bundles that work together across categorical boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If content items are grouped by categories or developer, then recommendations are organized systematically, but the recommendations fail to provide cohesive solutions that match user's unique interests and changing circumstances
Solution Approach 1:
The system uses co-selection scores derived from user behavior data to dynamically adjust and personalize recommendations. By analyzing which content items users download together, the system creates feedback loops that continuously refine recommendation bundles, making them adaptable to individual user preferences while maintaining systematic organization through the graph structure.
Solution Approach 2:
The patent segments the recommendation problem into two layers: a systematic layer (graph structure organizing content items by co-selection relationships) and a personalized layer (customized bundles selected based on user-specific co-selection scores). This segmentation allows the system to maintain organizational structure while delivering personalized recommendations.
2Measurement precision
If the system tracks and analyzes user download decisions to improve recommendation accuracy, then recommendation relevance improves, but the computational complexity and data processing requirements increase
Solution Approach 1:
The system calculates co-selection scores for all pairs of content items (excessive action) to ensure comprehensive coverage and high measurement precision. By pre-computing these scores for all possible pairs and storing them in the graph structure, the system pays the computational cost upfront rather than calculating scores on-demand, reducing real-time computational complexity when generating recommendations.
Solution Approach 2:
The system performs preliminary analysis of user download patterns to establish co-selection scores before actual recommendation generation. By pre-processing the data to create the co-selection graph and scores, the system eliminates the need for complex real-time calculations during recommendation delivery, thereby reducing operational computational complexity while maintaining high accuracy.
Data Source
AI summary
A system and method of recommending a bundle of content items to a user, including storing a plurality of content items in a computer system, determining a respective co-selection score for each pair of content items among the plurality of content items, the co-selection score indicating a probability that a given pair of content items among the plurality of content items will both be downloaded by a user of the computer system, and outputting, to a first user, a plurality of content items comprising a sub-set of the plurality of content items.