An Interactive Algorithm Research on Ontology-Based Association Rules Mining

摘要

In order to get more knowledge during the process of association rule mining, users have to deal with huge amount of rules, even though most of the rules are common sense. An Interactive Association Rules Mining System (IF-Apriori) is proposed in this paper. Through IF-Apriori, on one hand, users can use rule schemas to express their prior knowledge; on the other hand, the rule schema level can be defined by users according to their subjective interestingness in the system. After that, an interactive framework is proposed. Users can define new rule schemas and modify rule schemas level in every interactive process. Experimental results show that the number of rules is reduced and quality of the rules is improved with IF-Apriori.