An Effective Analysis Method of Discussions in Bulletin Board Sites

As both Internet environments and computer environments dramatically grow, it is easier and easier to discuss many topics with many people. Large amount of discussions are daily performed and stored in bulletin board sites. Most of the discussions are unimportant for most of people. But small proportion of the discussions may be able to have important influence on organizations and the companies referred to these discussions. If the organizations and the companies do not early recognize the discussions and do not timely perform appropriate countermeasures, their brand images may be injured or they may overlook the symptom of the boom leading to the sales. They are asked to early discover the discussions. However, since there are many bulletin board sites and many discussions are performed in each site, it is impracticable for the organizations and the companies to manually check all discussions in detail. On the other hand, they have various interests. Even if some discussions are important for an organization, the discussions are not always important for the other organization. The organizations and the companies need to discover discussions corresponding to their interests. They need a method that can easily analyze the discussions and can help them to discover true important discussions. This article introduces one of methods. The method extracts events related to their interests from discussions. The discussions are ranked by referring to the event. The discussions with high order are extracted as important discussions. The method also extracts keywords representing the discussions. The extracted discussions and their keywords help the organizations and the companies to quickly understand the topics of the discussions. The organizations and the companies can discover true important discussions without heavy burden. In the following sections, the method is introduced. BACKGROUND