The amount of information is exponentially increasing over the internet where information sources are autonomic, heterogeneous and dynamic. Thus, effective selective dissemination of relevant information is very essential for users. Publish/subscribe systems are gaining increasing popularity due to its loose coupling and scalability. However, traditional content-based pub/sub systems are crisp and limited. We put forward a new subscription model which could describe the users' fuzzy needs more precisely with similarity threshold and weight of each constraint. We also propose a semantic matching algorithm, which makes use of domain knowledge, described by fuzzy ontology, to solve the uncertainty problems. The experiment demonstrates the algorithm's high efficiency, high effectiveness and good scalability.
展开▼