Trend-Aware Entity Relationship Extraction for Recommendation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current recommendation systems in network-based publishing systems face challenges in providing accurate and relevant recommendations due to the randomness of browsing behavior, which is less structured than purchase data, and struggle to adapt to new trends and products, leading to misadvised suggestions and difficulty in generalizing across broad categories.
Innovation Solution
The implementation of trend-aware self-correcting entity relationship extraction methods that analyze user navigation patterns to create weighted entity relationship graphs, adapting to community behavioral trends and using edge strengths and probabilities to generate recommendations, allowing the system to rapidly adapt to new products and trends.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If collaborative filtering is used to generate recommendations, then recommendations can be produced based on user behavior patterns, but the system cannot adapt quickly to new trends and products
Solution Approach 1:
The patent implements dynamic entity relationship graphs that continuously evolve as new entities and relationships are discovered through user navigation. The system updates edge strengths and probabilities in real-time, allowing the recommendation structure to adapt dynamically to new trends while maintaining reliability through probabilistic reasoning and trend awareness mechanisms.
Solution Approach 2:
The system incorporates feedback loops where user navigation patterns are continuously monitored, entity relationships are extracted and updated, and recommendations are adjusted based on observed trends. This feedback mechanism enables the system to learn from user behavior and improve recommendation accuracy while adapting to emerging trends.
2Reliability
If purchase data is used for recommendations, then structured data provides reliable patterns, but browsing behavior data is too random and unstructured
Solution Approach 1:
The patent introduces entity relationship graphs as an intermediary structure that bridges structured purchase data and unstructured browsing behavior. By extracting relationships between entities from navigation patterns and representing them as weighted graphs with edge strengths, the system transforms random browsing data into structured, analyzable patterns while preserving the versatility of covering diverse user behaviors.
3Productivity
If traditional recommendation systems are implemented, then they can provide basic recommendations, but they produce misadvised suggestions and struggle with generalization across categories
Solution Approach 1:
The patent changes key parameters by introducing edge strengths and probabilities as dynamic weights in the recommendation system. Instead of treating all relationships equally, the system varies the importance of different entity relationships based on observed navigation patterns and trend data, enabling more precise and relevant recommendations while maintaining high productivity through efficient graph-based computation.
Data Source
AI summary
Methods and systems for trend aware self-correcting entity relationship extraction are disclosed. For example, a method can include receiving a selected entity, determining a plurality of entities related to the selected entity, determining a plurality of most probable entities, calculating relevance scores, and displaying a subset of the plurality of most probable entities. The selected entity can be received on a network-based transaction system. The plurality of entities related to the selected entity can be determined based on a relationship score. The relationship score can represent navigation transitions, aggregated over time, between the selected entity and each of the plurality of entities. The plurality of most probable entities can be determined based on probabilities. Relevance scores can be calculated for each of the plurality of most probable entities. Finally, the subset of the plurality of most probable entities to be displayed can be determined according to the relevance scores.