In this paper, we propose a new approach to measure sim-ilarities among academic papers based on their references. Our similarity measure uses both in-link and out-link by transforming in-link and out-link into undirected links.

A blog network, composed of blogs and their relations, may exhibit two different propensities characterized by the purpose of use: an information-oriented propensity and a friendship-oriented propensity. Both propensities coexist in a blog network, and the degree of these propensities may play an important role in business and policy decisions of blog-related business. In this paper, we propose an automated method for determining the propensity values of a blog network. First, classification is used to judge the propensity values of the relation between two blogs. Then, by adding up the propensity values of all the relations in the network, one determines the propensity values of the whole network. Through extensive experiments using a large volume of real-world blog data, we demonstrate our method achieves a high level of accuracy in determining the propensity values of a relation. The results also suggest the applicability of our approach for determining the propensity values of a network.

In the blogosphere, there exist posts relevant to a particular subject and blogs that show interests in the subject. In this paper, we define a set of such posts and blogs as "blog community" and propose a method for extracting the blog community associated with a particular subject. The proposed method is based on the idea that the blogs who have performed actions to the posts of a particular subject are the ones that have interests in the subject, and that the posts which have received actions from such blogs are the ones that contain the subject. The proposed method selects a small number of seed posts that contain the subject. Then, it selects the blogs that perform actions to the seed posts over some threshold and the posts that have received actions over some threshold. By repeating these two steps, it gradually expands the blog community. The experimental results show that the proposed method exhibits a higher level of accuracy than the methods proposed in prior research.