Rise and decline process of online communities : modeling social balance of participants

Abstract

Some online communities like Friendster had declined, and some of the others are said to be declining. Recent research has revealed the mechanism of decline as well as that of rise in each community. However, no comprehensive research has yet revealed the difference in declining mechanisms of each communities. We considered the online communities as networks of users and topics and defined behavior of users using Heider's balance theory. Users in our model are in a dilemma, stuck between topic preference and the balance between neighboring users. How the user behaves in the dilemma, his/her strategy, disseminates to other users. We simulate online communities using the model and observe the rise and decline of different kinds of communities. As a result, we found that two types of communities tend to develop with many users: communities in which the topic changes dynamically (FreeTopic-type) and communities in which the topic changes gradually (Topic-type). However, the property of each community and behavior of users are different. We found by simulation that the collaborative behavior of users happens very frequently in the FreeTopictype community, in which users consider the balance between each other rather than their topic preference. As a result, the FreeTopic-type communities do not often crash (i.e. quickly lose users). In addition, we confirmed that the postings about a topic are either negative or positive in the FreeTopic-type community. On the other hand, in the Topic-type community, simulation results indicate that users prioritize their preference for a topic. This causes the community to crash very frequently. However, users in such a community are found to obtain more benefits than in FreeTopic-type communities. It can be said that, after crashes occur, the community is still relatively beneficial for some users who remain.