Blogger clustering by utilizing link information

Blogs are a new form of internet phenomenon and a vast ever-increasing information resource, which are dated unedited, highly opinionated personal online commentary including all kinds of hyperlinks such as citation link, comment link, blogroll link. These links can be viewed as the blogger´s browse behavior, which reflects the user´s interest to a certain extent. So we construct a blogger-post matrix, link analysis is considered in calculation of the entry of the matrix. With usage of probability latent semantic analysis, the conditional probability of latent variable Z to post P is transformed the the conditional probability of latent variable Z to post B, then the transformed results are used in similarity calculation. The k-medoids algorithm is adopted to further improve clustering result. Experiment results have shown that this new algorithm is effective.