Evaluation of access patterns from log Ontology with DL-safe rules
Zhou Ming-tian · Journal of Computer Applications · 2009
In order to get the useful patterns from Web usage records,the author combined log Ontology and application access rules into a decidable hybrid log knowledge base with the restriction of DL-safe rules,and presented an approach for evaluating useful patterns from the candidate user-access pattern set.This method adopted inductive logic programming for coverage test about observations,and computed the support of user-access patterns to discover the frequent ones.The semantic generality measure could improve the results of patterns evaluation,and refining the random patterns by events taxonomy in log Ontology could improve the efficiency of patterns evaluation.Experimental results show that this method is feasible and effective to solve practical problems.