引文内容视角下的引文网络知识流动网络分析

The Effect of Knowledge Flow in Citation Network from the Perspective of Citation Content

【目的/意义】从引文内容视角探究引文网络的拓扑结构及其组成要素的知识属性与引文网络中知识流动效应间的关系。【方法/过程】首先,在阐述引文内容视角下引文网络知识流动主要特征的基础上,从静态特征属性和动态结构属性2个方面构建了知识流动的演化动力模型;然后,依据影响引文网络中知识流动效应的主要因素,从知识吸收效应、知识叠加效应和知识成本效应3个方面构建了知识流动效应函数的数学模型;最后,借鉴描述知识溢出效应的蜂巢模型及其修正模型提出了知识流动效应系数的计算方法。【结果/结论】运用Netlogo设置演化动力模型的初始参数,并通过仿真实验剖析各因素对知识流动效应系数的影响机理。文中方法可为客观反映引文网络中的知识流动规律与模式,并揭示引文网络中的知识流动效应提供一套可参考的实证模型。【创新/局限】本文通过计算实验方法从理论上较为系统地分析了引文内容视角下引文网络的知识流动效应,得出了一系列结论,但并未结合实际题录数据的计算结果进行对比和优化。

【Purpose/significance】The research aims to explore the relationship between the topological structure and the knowledge attribute of citation network and the effect of knowledge flow in citation network from the perspective of citation content.【Method/process】Firstly, on the basis of characteristics of knowledge flow in citation network from the perspective of citation content, the evolutionary dynamic model of knowledge flow is constructed from two aspects: static characteristic attribute and dynamic structure attribute;Secondly, according to main factors influencing the effect of knowledge flow in citation network, the mathematical model of the effect coefficient of knowledge flow is constructed from three aspects: knowledge absorption, knowledge superposition and knowledge cost;Finally, the method to calculate the effect coefficient of knowledge flow is proposed by using the beehive model and its modified model to describe knowledge spillover effect.【Result/conclusion】Netlogo is used to set the initial parameters of the evolutionary dynamic model, and the influence mechanism of each factor on the effect coefficient of knowledge flow is analyzed through simulation experiments.The method in this paper can provide an empirical model for objectively reflecting the law and pattern of knowledge flow in citation network and revealing the effect of knowledge flow in citation network.【Innovation/limitation】This paper systematically analyzes the knowledge flow effect of the citation network from the perspective of citation content through computational experiment methods,and draws a series of conclusions, but it does not compare and optimize the calculation results of actual bibliographic data.