Abstract
Ontologies are used to represent data and share knowledge of a specific domain, and in recent years they tend to be used in many applications such as database integration, peer-to-peer systems, e-commerce, semantic web services, bioinformatics, or social networks. Feeding ontological domain knowledge into those applications has proven to increase flexibility and inter-operability and interpretability of data and knowledge. As more data is gathered/generated by those applications, it becomes important to analyze and transform it to meaningful information. One possibility is to use data mining techniques to extract patterns from those large amounts of data. One challenging general problem in mining ontological data is taking into account not only domain concepts, properties and instances, but also hierarchical structures of those concepts and properties. In this paper, the authors research the specific problem of extracting ontology-based sequential patterns.
TopSemantic Web is designed to let users and machines describe resources, share that data in a distributed manner and enable interpretation and processing of the related data (Lee, Hendler, & Lassila, 2001; Lee, 2001). In recent years, ontologies are widely used to realize the data layer of Semantic Web. In this paper the specific problem of mining onlogy-powered systems is addressed. One reason for focusing in the knowledge discovery aspect of a Semantic Web environment is that the concepts and relations of an ontology have an impact on the quantity and quality of the patterns that may be extracted from such graph-based data (Di-Jorio, Bringay, Fiot, Laurent, & Teisseire, 2008; Liao, Chen, & Hsu, 2009; Rajapaksha & Kodagoda, 2008).
Mining data in the context of ontology-powered systems involves different fields such as ontology engineering, knowledge discovery and pattern mining. Hereafter, the background knowledge related to this study on both pattern mining and ontologies is presented.