Entity Tag Association Prediction Using Graph Convolutional Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing annotation systems for multi-tag classification problems are often incomplete due to the vast scale of tag libraries and sample datasets, leading to inaccurate descriptions and impacting real-life applications such as personalized recommendations.
Innovation Solution
An entity tag association prediction method using a graph convolutional network (GCN) is employed to construct entity similarity graphs, tag similarity graphs, and entity tag association bipartite graphs, enabling the prediction of missing tag associations and enhancing annotation completeness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large-scale tag library is used to provide fine-grained description, then the description accuracy is improved, but the annotation completeness deteriorates due to the immense scale making comprehensive annotation infeasible
Solution Approach 1:
The patent performs preliminary actions by constructing entity similarity graphs and tag similarity graphs before the main prediction task. These pre-computed graphs capture relationships between entities and tags, enabling the system to predict missing tag associations without requiring complete manual annotation of all possible tag-entity pairs, thus resolving the contradiction between comprehensive coverage and annotation feasibility
Solution Approach 2:
The patent introduces similarity graphs as intermediary structures that mediate between the large tag library and the limited annotation data. The entity similarity graph and tag similarity graph serve as intermediaries that encode relational information, allowing the prediction model to infer tag associations indirectly through these intermediate representations rather than requiring direct complete annotations
2Quantity of substance
If existing annotation data is used for multi-tag classification, then the training data availability is improved, but the prediction accuracy deteriorates due to highly incomplete annotations with numerous omissions
Solution Approach 1:
The patent implements feedback mechanisms where the prediction model continuously refines its accuracy by leveraging the similarity graphs. The model uses the pre-computed entity and tag similarity information to correct and enhance the incomplete annotations, creating a feedback loop that improves prediction accuracy progressively without requiring complete training data
Solution Approach 2:
The patent changes the parameter representation from direct tag associations to similarity-based embeddings. By transforming the annotation data into similarity graph representations, the system can extract meaningful patterns from incomplete data, improving prediction accuracy through parameter transformation rather than relying on complete original annotations
Data Source
AI summary
The present disclosure provides an entity tag association prediction method, device, system, and a computer readable storage medium. The method includes: determining an entity relationship network, a tag relationship network and an entity tag association network; constructing an entity similarity graph according to the entity relationship network, constructing a tag similarity graph according to the tag relationship network and the entity tag association network, and constructing an entity tag association bipartite graph according to the entity tag association network; extracting an entity feature, and constructing a tag feature according to the tag similarity graph; integrating the entity similarity graph, the tag similarity graph, and the entity tag association bipartite graph into a graph convolutional network to construct a prediction model; inputting the entity feature and the tag feature into the prediction model for training until the model converges, and outputting a prediction result of the prediction model.