Analysis And Research On User Behavior Based On Linear Threshold Model In Online Social Network

With the rapid development of social media(web media), users have more diversified ways to obtain information, as the main publishers of information users, by publishing interested in certain aspects of information users to establish a social link, conducting direct interaction between users and build their own social networks, online social network. Researching behavior characteristics of users in online social network, can help to find the social phenomenon or intrinsic value which is not easy to find in real life.Reference is presented in this thesis, on the basis of a large number of domestic and foreign research literature, analysis and study on the linear threshold model of online social network user behavior affect the application of the online social network data sets of collection and processing, using software programming to carry out a series of work, the main research contents and conclusions are as follows:In this thesis, YouTube, Flickr, and Twitter content of three different types of online social network relations of property of the publisher and the audience were analyzed, and excavating in three different media online social network behavior characteristics of the differences and similarities.First analyzing the general situation of the publisher and the audience, and then respectively analyzing the publisher and the characteristics of the audience, the second for publishers and audience and researching the interaction between them, finally, the intersection of the publisher and the audience, dual role, user behavior characteristics of analysis.The homogeneity is obtained by the analysis the applicability of the theory on the online social network, as well as the different information carrier of information transmission speed, breadth, and the influence of the life cycle, etc.Found online social network users have obvious homogeneity on three attributes in age, gender and location, namely the audience tend to choose with similar properties of publishers to upload content; At the same time, three kinds of dual role phenomenon common in network storage is existed, but the user as a different role is in the right activity, etc.As an online social network user behavior research and based on the discovery and the role of public opinion research as a starting point, use different means to collect and pretreat OSN of data sets.For the study of online social networks based on user behavior found that the hot issues of online social network user behavior modeling problem is also in the online social network under the new environment, exploring the new problems in the user’s behavior.A new generation of online social network provides more detailed user data, based on user behavior of online social network node role found that provides an important basis, and classifying these detailed behavior data for analysis, has important application value.The development of mobile Internet, micro blogging systems, the bits and pieces, timeliness of information of online social network group events can reflect the characteristics of network public opinion.Analyzed the key factors affecting the user clicks to users in Twitter post short click-through rates on behalf of its influence, and according to the analysis to predict the influence to the users.Research selects web sites for Twitter users and related data, for the purpose of analysis from three aspects to hit the key factors affecting users: attributes of users, the behavior of user and themes.Finally based on the above analysis to extract the corresponding feature,we use the prediction model to predict the influence of the user.In this thesis, the application of linear threshold model analysis framework in the influence of online social network user behavior, the technology method for personalized information recommendation, network management, advertising their society have certain guiding significance and practical value.