知识图谱内容关系重构与挖掘研究

Content Relation Reconstruction and Mining for Knowledge Graphs

摘要: 知识图谱的挖掘可以帮助发现新的知识，通过挖掘数据之间的隐含关系和规律，揭示事物的本质和内在联系。本文构建了植物与土壤关系的知识图谱，通过知识关系将图谱结构进行动态演化，然后对演化后网络的度、介数、接近度分布特征进行分析，并根据网络中三阶和四阶结构体分布特征识别出了高粱、梨、莴苣等关键的知识实体，可以反映出这些植物实体在植物与土壤关系方面知识应用的重要性和影响力，为基于知识图谱的智能问答设计提供参考。

Abstract: Knowledge graph mining can help discover new knowledge, by mining the hidden relationships and patterns among data, and revealing the essence and intrinsic connections of things. This paper constructs a knowledge graph of plant-soil relationships, and dynamically evolves the graph structure through knowledge relations. Then, it analyzes the distribution characteristics of degree, betweenness, and closeness of the evolved network, and identifies the key knowledge entities such as sorghum, pear, and lettuce based on the distribution characteristics of three-order and four-order structures in the network. These plant entities can reflect their importance and influence in the knowledge application of plant-soil relationships, and provide reference for the intelligent question answering design based on the knowledge graph.