On The Structure Of Wikipedia Knowledge Link Network And Patterns Of Its Editing Behavior

Knowledge is vital to the development of society and economy.Understanding the regular patterns of development of large-scale knowledge networks is of great significance to the innovation and growth of knowledge.As one of the largest knowledge communities in the world,Wikipedia is of great significance to the development of knowledge.The difference between traditional encyclopedias and Wikipedia is that the former is written by experts,while the latter is edited by a large number of volunteers.This kind of self-organizing collective knowledge building has attracted the attention of many researchers.With the explosive growth of knowledge,self-organized collective knowledge building will become one of the most important ways of knowledge integration.Understanding the knowledge building mechanism in Wikipedia is of great importance for the development of knowledge.The link structure of Wikipedia affects the efficiency of knowledge retrieval and the development of Wikipedia.However,the existing research lacks analysis of the formation of Wikipedia's link network.To address this issue,we used items in Wikipedia biology category as an example,clearly showed the structural characteristics and evolution process of the knowledge link network from a micro perspective.The migration mode of people in the knowledge space is not only closely related to the formation of this network,but also an important part of collective knowledge building.The existing researches on the behavior of editors always cut off the association between editor's behavior and knowledge structure,or ignored people's subjective initiative in the process of knowledge building.This thesis quantitatively portrayed the editor's subjective editing tendency,and attempted to show the role of different people in the process of collective knowledge building by analyzing the mode of transfer and recency of editors' behavior.(1)Through the analysis of the structure and evolution of the knowledge link network under the Wikipedia biology category,we intuitively displayed the process of collective knowledge building.This knowledge link network is a highly modularized small-world network.This thesis summarized how different knowledge modules connected by analyzing the role of nodes and the structure of the diverse-club.The skeleton of the network is composed of hub nodes and diverse-club nodes,other nodes with low-degree and low participation coefficient are filled in the network,which affects the scale of the network.Some knowledge clusters appeared in the network at first.These clusters with star structuresdeveloped independently in the early stage,and then gradually formed a complex network with a good structure through the bridging of some diverse-club nodes.This kind of knowledge integration that expands from local to global leads to a sparse but efficient small-world network.This thesis clearly described the specific structure of the Wikipedia link network that has not been shown in previous studies.The formation of this network was taken as an example to show the efficient mode of collective knowledge building.It was found that human knowledge organization is similar to the structure of brain network.This thesis clearly demonstrated the process from decentralization to aggregation of the knowledge link network,which can not only help people understand the high-efficiency patterns of knowledge integration,but also has important reference significance for other knowledge communities.(2)This thesis quantitatively analyzed the editor's transfer mode in the knowledge space from a new perspective by introducing the knowledge link network as a map to analyze the patterns of editor's interest transfer.We also analyzed three factors which affect the editor's transfer behavior: heterogeneity of editing ability,topic similarity editing tendency and recency editing tendency.It was found that editors who edit more items tended to transfer more frequency,and they mostly move between items with similar themes.In addition,by analyzing the effect of recency in editing behavior,we found that editors are very concerned about the items which they edited recently,and the editors' editing abilities in different items are different.This heterogeneous ability and recency editing behavior may affect each other,prompting the editors to return to the items which edited recently.This tendency can help to improve the quality of the items through repeated editing.This thesis summarized the similarities and differences between the behavior of robots and human editors.Although robots have similar patterns of interest transfer as human editors,its' probability of transfer in a small range is much higher.Besides,robots do not like human editors to have a strong tendency to return to the items they edited recently.This difference may because the robots' editing behavior are mostly to maintain Wikipedia,while human editors are more concerned about the contents they edited.They work together to promote the development of Wikipedia and the evolution of the item link knowledge network.Our research showed the different transfer patterns of different editors,which can help to coordinate and give full play to the role of different groups in the process of knowledge building,which has important reference significance for improving the efficiency of knowledge building.