A New Assessment Method On Impact Of Node In Sina Microblog Community

Chinese microblog is an important way of social network and social mediaexchange,and is the most typical applications of Web2.0era.In August2009,Sinamicroblog developed into China’s mainstream microblog, it is posted by instantcommunication no longer than140words.With the development science andtechnology, People can publish their own information through mobile phones, pad,API interface.According to CNNIC statistics in2011, Sina number of users hasexceeded300million. In the the microblog complex social network, we mainly studyuser behavior and relationships of the bloggers. We are concerned about those whohave a strong influence bloggers, and have some further analysis of other users whohave relationship with these high-impact bloggers, Usally users who have morenumber of fans have more influence.Because of commercial interests, the situationhas changed,Traditional PageRank algorithm is a page sorting algorithm based on thenumber of Web links, We sort bloggers influence by this algorithm, But result is notsatisfactory. This article is based on user network and user interaction, we design anew node impact assessment method--User Impact Rank algorithm.In this paper, we consider the traditional sorting algorithm, and sum up thefollowing three questions.First, the discrimination of the new user.The number of users is increasingrapidly,microblog network becomes more complex, so the user impact assessmentalso become more difficult. Old users join the microblog community longer time thanthe new users, their fans and microblog information are more than those new users’,their microblog information is more likely to cause interaction between users thanthose new users’microblog information, But Some of the new user’s influence issignificantly higher than the old users’. Traditional sorting algorithm does not take thisinto account.Secondly, the number of fans causes a user’s qualitative change. The number ofusers’ fans is an important criterion, but not the only standard to measure the influenceof users. The number of two users’fans is very diffrent, but their influence is the sameby the traditional sorting algorithms. In Fact we hope users who have fewer fans havemore influence, because these users have more qualitative fans.Thirdly, water forces and zombie fans interfere with the sorting.In general, Ifblogger have a lot more fans,he will be more influencial.However, The Chinesemicroblog has two unique phenomenon:(1)Many celebrities buy a lot of zombie fansto increase their own influence.(2)Some teams build up a huge machine account to form artificial diffusion and thermal area, these accounts are the water forces.Falseuser and false interaction lead to the inaccurate sort of influence by users’ fans.Our research is based on user networks and user interaction behavior in Sinamicroblog community. This paper describes the principle of traditional PageRank,Behavior-Relationship Rank and User Impact Rank algorithm, we have conducted acomparative experiment and prove that UIR algorithm is more objective and accurateto assess the influence of the users’ nodes.