Ontological Knowledge Inferring Approach based on Term-Clustering and Intra-Cluster Permutations

Abstract

Ontological representation of knowledge has the advantage of being easy to reason with, but ontology construction with knowledge facts, automatically acquiring them from open domain text is often challenging. This research introduces a novel approach to infer new ontological knowledge in a fully automated manner. Such ontological knowledge can be utilized in both constructing new ontologies and extending existing ontologies. Basic level triples that can be extracted from open domain text are used as the data source for this study. A simple mechanism has been introduced to convert the triple into an ontological knowledge fact and such ontological knowledge facts are further processed to infer new ontological knowledge. The main focus of this research is to infer new ontological knowledge using an advanced term-clustering mechanism followed by an intra-cluster permutation generation task. Generated permutations are potential to be selected as good ontological knowledge facts. Inferred ontological knowledge was tested with inter-rater agreement method with high reliability and variability. Results demonstrated that, out of 43,103 triples, this method inferred 127,874 ontological knowledge (approximately 3 times) of which 66% were estimated to be effective. Finally, this research contributes a reliable approach which requires a single pass over the corpus of triples to infer a large number of ontological knowledge facts that can be used to construct/extend ontologies.