Knowledge Graph Relationship Prediction Model for Intelligent Search

Intelligent search system can effectively realize the integration of multi-source heterogeneous data resources. In the intelligent search and question answering system based on knowledge graph, the questions are usually asked by some simple questions, which are processed in a pipeline way, and then the subject entity and predicate relationship of the question are selected, and then the most likely answer entity is selected by information integration. Entity relation extraction is the difficulty of knowledge graph. We found that in traditional information processing models, neural network models usually only utilize question sentences for relation classification. But such question sentences often lack context, making it difficult to fully utilize the computational power of neural networks. In response to the above problems, this paper proposes a knowledge graph relationship prediction model (BXS) for intelligent search systems, which uses sentence subject headings and entity background information as auxiliary reasoning information to obtain sufficient context for the problem. We utilize two attention mechanism models to fuse auxiliary information with questions for semantic matching. The experimental results on the data set prove that the model can effectively improve the accuracy of relation extraction, and effectively improve the accuracy of the final question and answer.