基于图的文献引用关系挖掘算法的研究

Research on Mining Algorithm of Graph-based Literature Citation Relationship

基于图结构的挖掘是数据挖掘的新的研究方向，根据相关内容定义了用图表示两个有关文献引用关系的概念，提出了用图挖掘解决文献间相似关系的衡量与解决方法，目的在于找寻以文献为结点构造的图形中结点和结点之间的相互关系，进而反映出文献和文献之间的相似关系与文献的权威度。将图挖掘技术应用在文献引用关系方面，并提出了比较新颖的挖掘算法，此算法适合于以文献为结点构造的有向无环图，并为将来继续研究此方向打下了基础。所得结论将在图挖掘理论研究和网络搜索引擎方面具有很高的实用价值。

The Mining based on graph structure is the new research direction of data mining. According to the related content, the paper defines two relevant citation relations, and based on graph mining, proposes a method of the measurement and solutions of the relation ship between similar papers. What this paper＇s purpose is to look for the interrelation between nodes and nodes in the graph where references is abstract to nodes, and then reflect the similar relations among the references and the authority of the reference. The technology of Graph Mining is applied in reference,and this paper＇s novel algorithm for Graph Mining which is suitable for the Directed Acyclic Graph, lays the foundation for continuing studying in the direction. The conclusions of the paper in research for theory of the Graph Mining and Search Engine have the strong practical value.