SWApriori : a new approach to mining association rules from semantic web data

With the introduction and standardization of the semantic web as the third generation of the Web, this technology has attracted and received more human attention than ever and thus the amount of semantic web data is constantly growing. These semantic web data are a rich source of useful knowledge for feeding data mining techniques. Semantic web data have some complexities, such as the heterogeneous structure of the data, the lack of exactly defined transactions, the existence of typed relationships between entities etc. One of the data mining techniques is association rule mining, the goal of which is to find interesting rules based on frequent item-sets. In this paper we propose a novel method that considers the complex nature of semantic web data and, without end-user involvement and any data conversion to traditional forms, mines association rules directly from semantic web datasetsat the instance level. This method assumes that data have been stored in triple format (Subject, Predicate, and Object) in a single dataset. For evaluation purposes the proposed method has been applied to a drugs dataset that experiments results show the ability of the proposed algorithm in mining ARs from semantic web data without end-user involvement.