Profiling of Participants in Online-Community

Promoting interactions among participants in an onlinecommunity is catching attention of web sites’ managers. In this paper, we first introduce Influence Diffusion Model (IDM), a method for discovering influential comments, participants and terms from threaded online discussions, and evaluate the performance by precision and recall measurement. Then we propose a new method for profiling of participants in an onlinecommunity by expanding the idea of IDM. The positioning maps derived from the profiles show the relations among participants as well as their characteristics. Introduction Communication places on the Internet, such as BBS, chat room etc. are designed to gather people into web sites by promoting the interactions among participants (Ishikawa 2001). Previously we had proposed Influence Diffusion Model (IDM) (Matsumura, Ohsawa, & Ishizuka 2002) which can discover influential participants, comments and terms from threaded online discussions. In this paper, we propose a method for profiling of participants in an onlinecommunity by expanding the idea of IDM. Here, we regard a set of influential terms of a participant as his/her profile. The profiles help us understand their characteristics by which we can manage the participants for aiding their interaction in the community. Influence Diffusion Model Firstly, we introduce IDM which is designed to measure the influence of comments, participants and terms by the degree of text-based relevance of comments under the situation that interactions among participants are done by exchanging comments, i.e., posting new comments or replying to the comments. The threaded comments, called comment-chain, show the flow of influence. For example, if a comment replies to a comment , it is considered that is affected by . Similarly, if a participant replies to a comment of a participant , is considered to be affected by . In these cases, the influence diffuses from to / from to . Here, IDM defines the process of diffusion of influence as follows. Copyright c 2002, American Association for Artificial Intelligence (www.aaai.org). All rights reserved. Definition 1 In text-based communication, influence diffuses along the comment-chain by medium of terms, i.e., words or phrases. On the basis of Definition 1, the influence is defined by the degree of terms propagating throughout the comment-chain. For example, If replies to , the influence of onto , , is defined as