Font Size: a A A

Feature Analysis And Topic Evolution Of Dynamic Tagging Knowledge Network In Zhihu Community

Posted on:2022-11-18Degree:MasterType:Thesis
Country:ChinaCandidate:H S ZhangFull Text:PDF
GTID:2518306728472004Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Segments as mobile Internet applications and user information demand to upgrade,to question and answer community q&a session as the main characteristics of the network has gradually become the important of people getting information and disseminate knowledge platform,its internal knowledge network structure profoundly affects the knowledge dissemination and sharing of logic,shaping the topic evolution mechanism,determines the whole ecosystem development platform for knowledge.In the Q&A community of Zhihu,each knowledge resource is labeled with a corresponding topic tag.The tag knowledge network is an important basic paradigm for in-depth exploration of the knowledge transmission and evolution trend in the community,and its research is the premise for mining the relationship between knowledge.Based on complex network theory and time series analysis,this paper makes an in-depth study on the tag knowledge network of Zhihu community from the perspectives of global network and local nodes.On the one hand,from the perspective of the overall network research,the structural characteristics of tag knowledge network in different periods are deeply and dynamically studied from the four dimensions of network topology,network size,network degree and network clustering.On the other hand,from the perspective of local nodes,key topic tags in different periods are identified,and the evolution of their characteristics and communities is dynamically analyzed.The research results are mainly as follows: First,The label knowledge network of Zhihu has the basic nature of social network,showing "small-world phenomenon" and "scale-free characteristic",and the transmission performance and efficiency of knowledge resources in the label knowledge network show a trend of gradually improving and then weakening;Secondly,the scale and clustering evolution of tag knowledge network can be predicted by ARIMA(0,1,2)and ARIMA(0,2,1)models;Third,topics such as "life","film" and "law" all play an important role in the process of knowledge dissemination,whether they have direct influence or act as "middlemen",while topics such as "investigation questions" and "psychology" have more significant direct influence than their intermediary effect.In addition,The mediating effect of "Internet","game" and "music" is more significant than its direct influence;Fourth,the topic of "law" has a strong connection with the topic of "medicine",the topic of "education" and the topic of "psychology".By integrating the research results,a guiding and supporting conclusion can be drawn,which is conducive to understanding the internal logical relationship between zhihu q&A community labels,facilitating better organization and utilization of network knowledge resources,thus providing suggestions for promoting zhihu knowledge dissemination and controlling public opinion,and providing support for research in knowledge recommendation and other fields.
Keywords/Search Tags:Tag knowledge network, Network structure characteristics, Zhihu's topic, Network community, Time series analysis
PDF Full Text Request
Related items