| With the rapid development of Internet technology, people search and share experience through social networks. The individual role in social network refers to the social status, behavior or virtual identity of the individual who participate in the network activities. Recognition of the role of social networks is a crucial task. For example, those who have the "expert role" of the users can provide a more accurate answer to the question in a technical forum. The quality of search service and the quality of recommendation of users of social network can be effectively improved using the analysis and identification of individual roles.The system based on identification of the expert role and the identification of influencer based-on changes of opinion has been widely used in the major social networks in real life. To improve the existing role recognition methods, and improve the role of identification, researchers are constantly introducing more new identification information. However, current method identifying the role of experts and influencing role is mainly link- or activity-based in social network. This kind of algorithm can identify types of users who are more active or answer more questions in the network.At present, the network is full of a lot of advertising, spam, malicious information, etc. Link- or activity-based method can’t identify the role of non expert who publish a large amount of garbage information or promotion of advertising in network, which may release users who publish a large amount of garbage information or promotion of advertising as experts. Therefore, the content of the exchange of information among the participants in the social network activities, as one of the characteristics of the participants’ role, has become an important part of the identification of the experts’ role.In this paper, we discuss the method of using the text quality analysis method to identify the expert role algorithm, and put forward the method of combining the text quality analysis with the expert role recognition algorithm based on link or activity method, the specific work is as follows:(1) The method of the expert role identification based on text quality analysis algorithm is presented. And comparing with manual identification expert users, it’s confirmed that expert role and influences can be identified by the content of participants’ communication in social network. It can effectively remove users who publish substantial advertising, garbage information and so on. It can increase the credibility of the role of experts.(2) Combination of method based on the text quality analysis and method based on link or activity. Firstly, calculate the social network user’s degree of expert using method based on the analysis of the quality of the text, and then use the link- or activity-based method to calculate the social network user’s degree of expert. Two different methods will be used to get the user expert degree of linear fitting, the social network users to calculate the degree of comprehensive expertise sort, so as to identify the role of experts. Experimental results show that the quality analysis of the text and link- or activity-based operating method’s combination of expert sorting algorithm can find active experts who answer questions more, and exclude users who publish a large number of spam. The result is closer when comparing with manual sorting.The role difference of the information publisher determines the credibility and importance of the search results. For example, if the role of the information publisher is an expert, then the credibility of the information is higher, the search results ranking of the information should be enhanced. Again, if the information released by the role is the effect, the information has a high influence, should also improve the information ranking. For new and old users in social networks, a good list of existing users can provide users with a more rapid and reliable service, greatly enhance the quality of social networking services. |
PDF Full Text Request