A Semantic Clustering Algorithm Oriented to Web Log

Existed Web logs are lack of semantics obviously. To improve the efficiency and accuracy of Web mining, a semantic Web log model - SWLM is presented, and two algorithms based on this model is given to cluster pages and users. Firstly, this method defines ontologies to interpret properties of Web log, then computing the semantic distances of log concepts. Based on the set of terms defined by ontology, the semantic information can be mining from the Web log. The test experiment shows that this model has better performance and clusters pages and users effectively. Those results can facilitate personalized services and user modeling