| With the development and application of web2.0 technology, virtual social networking community but also the rapid development of the Internet. Twitter, Facebook, Weibo, QQ, Wechat and other on-line social networking services as the main site for attracting billions of users on social development had a tremendous impact. Online social platform for mining, the focus is not only academic, but also had a huge economic and social value."Scientific network blog," real name system is a virtual community social networks, users are mainly researchers, university teachers and other senior intellectuals. Community members exchange is usually explore issues of frontier and innovative. Science blog network full network analysis helps to explore the basic relationship between bloggers and commentators, and analysis of community structure and node center will help you find the members of the community knowledge of communication and user behavior rule. Blog comments analysis can reveal the laws of blog authors and readers of the interactive behavior of the virtual environment to promote academic exchanges, to build a better online communication platform based on user interactions and foster the sharing of scientific knowledge and innovation, and promote academic exchange development activities.This article was first conducted for "scientific network blog" community more systematic research, "Science blog network," network analysis from the overall network architecture, node attributes, associations and other aspects of nature, the laws give virtual social behavior between bloggers; this article of blog analyzing information to obtain a behavioral pattern blog users and do correlation analysis for comment, get associated with comments feature, establish Bowen comments theoretical model based on such features, and the use of multiple linear regression of the theoretical model to predict; Bowen topic bloggers and social relationships have a greater impact on the commentary, this paper established a new regression model in multivariate linear regression models add Bowen themes and social relations two factors, the results show, added Bowen themes and social relationships higher regression model accuracy. |
