User Profile Generation via Entity Augmentation and Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current user profiling methods for content recommendation rely heavily on surface-level features and demographic data, failing to accurately capture specific user interests due to their general nature, which limits the effectiveness of personalized content suggestions.
Innovation Solution
A method that utilizes deep semantic knowledge by identifying and augmenting named entities from user queries and webpage content, clustering them into hierarchical structures to generate user profiles that reflect multiple aspects of user interests, thereby improving the accuracy of interest prediction and content recommendation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If surface-level features and demographic data are used for user profiling, then the system complexity is low and ease of operation is high, but the precision of user interest capture is insufficient
Solution Approach 1:
The patent transitions from surface-level feature analysis to deep semantic analysis by introducing a new dimension of entity-based profiling. Instead of relying on demographic data and browsing statistics, the system extracts named entities from user queries and webpage content, then augments these entities with semantic knowledge from knowledge graphs to create multi-dimensional user profiles that capture specific interests with high precision.
Solution Approach 2:
The patent introduces named entities as intermediary elements that bridge user activities and user interests. By extracting entities from queries and content, and then using these entities as keys to query knowledge graphs for semantic relationships, the system creates an intermediary layer that transforms raw browsing data into structured user interest profiles, resolving the contradiction between simplicity and precision.
2Measurement precision
If general topics are used to infer user interests, then the coverage is broad, but the specificity of interest prediction is insufficient
Solution Approach 1:
The patent segments user interests into specific named entities rather than treating them as general topics. By extracting individual entities from user queries and webpage content, and then creating separate profile entries for each entity with its semantic relationships, the system preserves specific entity information while organizing it into a structured format that maintains both granularity and comprehensiveness.
Solution Approach 2:
The patent creates composite user profiles by combining multiple augmented entities with their semantic relationships. Each user profile is a composite structure containing the original entity, augmented entities from knowledge graphs, semantic relationships, and user interaction data, which together provide both specific entity-level precision and broad topic coverage.
3Reliability
If deep semantic knowledge and entity augmentation are implemented, then the accuracy of content recommendation is improved, but the processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary action by pre-extracting named entities from user queries and webpage content during the browsing phase, and pre-augmenting these entities with semantic knowledge from knowledge graphs. This preliminary processing creates ready-to-use user profiles that can be quickly retrieved and applied for content recommendation, reducing the computational burden and processing time during the actual recommendation phase.
Data Source
AI summary
The present teaching relates to generating user profiles with semantic knowledge. A first information associated with a user is obtained. One or more entities are identified from the first information. The one or more entities are augmented based on second information to generate a set of augmented entities. The set of augmented entities are clustered into a set of hierarchical clusters. A set of user profiles is generated based on the set of hierarchical clusters so that the user profile is to be used to personalize content recommendation.