【摘要】 [目的/意义]针对大数据环境下用户兴趣数据稀疏、缺乏关联和描绘不准确等问题,利用知识图谱融合多源兴趣知识,以提高用户兴趣的全面性和准确性。[方法/过程]从兴趣之间的关联视角出发,进行兴趣建模、知识获取和知识融合,整合兴趣间的语义关联和社交网络关联,构建兴趣知识图谱;挖掘兴趣标签节点与上位词节点、百科标签节点、社交网络用户节点的关系,计算兴趣标签的语义关联度和社交网络关联度,生成复合关联权重,重构兴趣之间的衍生关系以实现用户的兴趣扩展。[结果/结论]该模型能够有效融合扩展不同类型的兴趣关联知识,相对于单一来源数据在用户兴趣的覆盖率和查准率方面均有所提升,提高了用户兴趣描绘的全面性和准确性。更多还原

【Abstract】 [Purpose/Significance] Interest data in big data environment is sparse,and there is no effective correlation in user interests. In response to these problems,a User Interest Expansion Model based on Knowledge Graph is proposed. [Method/Process] Starting from the perspective of the association relationship between interests,the model integrated the semantic associations and social network associations in interests through the process of interest modeling,knowledge acquisition and fusion utilization,and constructed an interest knowledge graph. The relationship among Interest Tag Nodes,Hyper Nodes,Encyclopedia Tag Nodes,and Social Network User Nodes was calculated,and the semantic relevance of interest tags and social network relevance were calculated to generate composite relevance weights. And the derivative relationships between interests were reconstructed to achieve user interest expansion. [Result/Conclusion] Experiments show that this model could effectively integrate different types of interest-related knowledge,and greatly improve the coverage and accuracy of user interest. It could improves the comprehensiveness and accuracy of user interest description.更多还原