Domain relation learning from events in log ontologies based on hierarchy association rules
基于层次关联规则的日志本体事件领域关系学习*

In order to discover the potential user-access patterns in Web usage records,this paper presented an approach for domain relation learning from events in log ontologies based on hierarchy association rules. This method fixed on the granularity of transactions through the part-whole relation between complex events and atom events in log ontologies, and extended the association rules mining algorithm on the hierarchy of events to discover the candidate frequent usage rules. Domain relations of events could be extracted after refining the redundant and noneffective rules. The experimental results show that this method can enrich the log ontologies and it is quite feasible and effective to solve practical problems.