Discovering Relevant Semantic Associations Based on User Specified Context

Abstract

Semantic associations are the complex relationships that exist between two entities in an RDF knowledge base. As the complexity and size of ontologies are increasing rapidly, the numbers of semantic associations are becoming increasingly overwhelming between a pair of entities. Hence, ranking of semantic associations is required in order to present more relevant semantic associations to the user. One of the criteria to find semantic associations is based on the context. Existing systems allow the user to define context by selecting concepts or regions from the ontology only at the level of schema but they do not take into account defining the context at the instance (or entity) level. As there exist a large number of instances for a schema concept in the knowledgebase, defining the context at the instance level may help the user in retrieving useful associations. In addition to this, present systems do not allow the user to retrieve the semantic associations based on his/her interested relationship though relationships are the important components of the Semantic Web. Due to this, sometimes user gets too many associations which require further investigation to get desired associations. To overcome these problems, this paper proposes a novel method to retrieve relevant semantic associations. Specifically, it proposes two parameters namely, Entity weight (which capture user’s interest at the instance level) and Relationship weight (which capture user’s interest on relationships) which are coupled with the context (defined at the schema level) can define user’s domain of interest more effectively thus produce user interested semantic associations. To make obvious the effectiveness of the proposed method, SWETO data set is used and the results demonstrate that the proposed method retrieves relevant semantic associations than the existing methods.