结合DL-safe规则的评估日志本体访问模式

Evaluation of access patterns from log Ontology with DL-safe rules

为从W eb使用记录中获得有效模式,在DL-Safe规则的限定下,将日志本体和应用访问规则相结合构建为一个推理过程可判定的混合日志知识库,并在此基础上提出一种从候选用户访问模式集中评估有效模式的方法。该方法首先借助归纳逻辑编程的思想对候选模式进行观察覆盖测试,并通过计算模式支持度找出其频繁项,然后利用语义普遍性测量提高了模式评估的质量。同时该方法还引入日志本体事件分类关系修剪冗余访问模式以提高模式评估的效率。实验结果表明了该方法的有效性和可行性。

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.