An ILP Approach to Mine the Association Rules on Log Ontology

Building rules on top of ontology is the main task of mining the logical layer of the semantic Web. In order to find the user-access patterns in the Web usage records, an ILP approach to mine the log ontology based on AL-log is illustrated in this paper. At first this paper gives the definition of event and log ontology, and then discusses the semantics of AL-log. Compared with traditional Web log files, log ontology has more expressive semantic information, and AL-log is a powerful hybrid knowledge representation and reasoning system by integrating description logics and Horn clause rule. Therefore applying AL-log to build the knowledge base of the log ontology can discover more effective and useful access patterns. For overcoming the "value restriction" of ALC, our ILP framework adopts a defined Datalog atom to express the part-whole relation between atom events and complex events, and then it makes use of ALC propagation rules and constrained SLD-reputation to learn the frequent association rules on log ontology. The experimental results show that this method can help site owners to create access rules effectively and it is quite feasible to solve practical problems.