The domain classification algorithm based on KNN in Micro-blog

The Influence of Micro-blog User is essentially the interactions of user to user. With the greater interactions of one user to another user, the influence of the user will be bigger. The existing researches about micro-blog influence are mainly aimed at Twitter and don't consider comment function, the characteristic of interdisciplinary and intersectional domain about the micro-blog. In order to fully consider those factors, we proposed the domain classification algorithm based on KNN in micro-blog. This algorithm classifies all micro-blogs into 15 types depend on the Open Directory Project. It determines the domain of miro-blog according to the degree of similarity between the content of micro-blog and domain ontology of Open Directory Project. Finally we get the influence of user in every domain by quantizing some factors, for example, the degree of similarity between the content of micro-blog and every domain, the number of fans, the number of retweeting, the number of commented, the login and registration time. The experiments show that the proposed method can get more accurate and practicability result than traditional ones, which fully considers the characteristic of interdisciplinary and intersectional domain.