Entity Page Recommendation via Post Content Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Social networks currently provide recommendations for following entities only based on interactions with post creators, failing to account for user interest in posts created by other users, leading to ineffective recommendations and a degraded user experience.
Innovation Solution
A system that detects user affinity by analyzing the content of posts, using a knowledge graph to identify related entity pages, and recommending them to users, even if the interaction is not directly with the entity, through a processor-executed method involving annotation, topic determination, and machine learning to ensure relevant recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If recommendations are provided only based on direct user-entity interactions, then the recommendation system is simple to implement, but the recommendation relevance and user engagement are insufficient
Solution Approach 1:
The patent introduces an intermediary mechanism (content analysis system with topic modeling and entity recognition) that mediates between user posts and entity recommendations. This intermediary analyzes post content to infer user interests and maps them to relevant entities, resolving the contradiction by adding complexity only where needed to improve recommendation relevance without overcomplicating the entire system.
Solution Approach 2:
The system performs preliminary actions by pre-analyzing and annotating post content with topics and entities before recommendations are generated. This advance processing creates a knowledge base that enables relevant recommendations without requiring complex real-time analysis, thus improving reliability while managing system complexity.
2Adaptability or versatility
If entity recommendations are provided only after user interaction with entity posts, then the recommendation logic is straightforward, but the system fails to capture user interest in content from other users
Solution Approach 1:
The patent implements a universal content analysis system that processes posts regardless of their source (user-generated or entity-generated). This multi-functional approach enables the system to capture user interests from any post content and generate appropriate entity recommendations, expanding recommendation coverage while using a unified analysis framework to manage complexity.
Solution Approach 2:
The system shifts from a single-dimension approach (direct user-entity interaction) to a multi-dimensional approach by analyzing content semantics, topics, and entities. This dimensional expansion allows the system to capture indirect user interests through content analysis, improving adaptability while using structured processing methods to control analysis complexity.
3Measurement precision
If content analysis and topic determination are performed for every post, then recommendation accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The system performs content analysis and topic determination as preliminary actions during post creation or caching, rather than in real-time during recommendation generation. This advance processing stores annotated content metadata that can be quickly retrieved for recommendations, improving topic identification accuracy while minimizing processing time impact on user experience.
Solution Approach 2:
The patent extracts only the essential elements (topics and key entities) from full post content through selective analysis. This extraction approach achieves sufficient topic identification accuracy by focusing on critical information rather than analyzing every detail, thereby reducing processing time and computational resources while maintaining recommendation quality.
Data Source
AI summary
Methods, systems, and computer programs are presented for an online community management system. One method includes detecting that a user has interacted with a post on a social network. The social network provides a plurality of entity contents. Further, the method includes determining a topic of the post based on content of the post. The method includes recommending following a related entity content of the plurality of entity contents to the user in response to determining that the topic is associated with the related entity content, the user is not following the related entity content, one or more users of the social network have posted in connection with the related entity content within a threshold amount of time, and the user is not blocked from accessing the related entity content.