论文部分内容阅读
[目的/意义]在保障知识网络的整体性征的条件下,从原始知识网络中提取具有显著意义的层次知识网络,奠定基于关联频度提取层次知识网络的理论基础。[方法/过程]以无标度网络与分形几何作为基础理论支撑,在对关键词知识网络和标签知识网络关联频度分布进行分析的基础上,采用关联频度作为阈值提取层次知识网络。并对层次知识网络的无标度性和小世界效应两项网络整体性征进行验证。[结果/结论]知识网络的关联频度分布服从幂律分布。以关联频度为阈值提取的层次知识网络在节点度值分布和关联频度分布方面都保持了原始整体网络的无标度性。层次知识网络能够很好地保持原始网络所具有的小世界特征。基于关联频度提取的层次知识网络与原始知识网络等效。
[Purpose / Significance] Under the condition of guaranteeing the overall character of knowledge network, we extract the hierarchical knowledge network with significant meaning from the original knowledge network, and lay the theoretical foundation for extracting the hierarchical knowledge network based on the association frequency. [Method / Process] With scale-free network and fractal geometry as the basic theoretical support, based on the analysis of the correlation frequency distribution of keyword knowledge network and label knowledge network, the association frequency is taken as the threshold to extract the hierarchical knowledge network. And verifies the global characteristics of scale-free and small-world effects of Hierarchical Knowledge Networks. [Results / Conclusions] The frequency of knowledge network distribution obeys power law distribution. Hierarchical knowledge networks extracted with the associated frequency as the threshold keep the original scale-free network in terms of node degree distribution and associated frequency distribution. Hierarchical knowledge network can well maintain the original small-world characteristics of the network. Hierarchical knowledge networks extracted based on association frequency are equivalent to original knowledge networks.