The goal of this project is to propose a novel method in mining association rules from semantic web data and linked data. Association rule mining techniques need transaction in order to mine association rules. But in semantic web data there is exact definition of transaction. All researches that had done on mining association rules from semantic web data so far, by user assistance launch to define transactions from semantic web data which this needs the user be familiar with semantic web structure and used data domain. Thus a new method is required to mine association rules from semantic web data without any need to transaction and also without end user involvement. For satisfying this requirement, a system had implemented in order to directly mine association rules from semantic web data, regardless to transaction concept and also without user involvement. This system consists three phases: generating 2-large itemset base on entities and their relations, generating larger itemsets and finally generating association rules base on large itemsets. This system also is able to mine association rules from a dataset that has generated from linked data concatenation. Regard to workflow and structure of this system, mining association rules from semantic web data, only demands a dataset that consists triples. The obtained results show that the proposed method, without user involvement can directly mine association rules from semantic web data and indirectly mine association rules from linked data.
