Enriching Ontology Concepts Based on Texts from WWW and Corpus

Tarek F. Gharib (King Abdulaziz University, Saudi Arabia)

Nagwa Badr (Ain Shams University, Egypt)

Shaimaa Haridy (Ain Shams University, Egypt)

Ajith Abraham (Machine Intelligence Research Labs (MIR Labs), USA)

Abstract: In spite of the growing of ontological engineering tools, ontology knowledge acquisition remains a highly manual, time-consuming and complex task. Automatic ontology learning is a well-established research field whose goal is to support the semi-automatic construction of ontologies starting from available digital resources (e.g., A corpus, web pages, dictionaries, semi-structured and structured sources) in order to reduce the time and effort in the ontology development process. This paper proposes an enhanced methodology for enriching Lexical Ontologies such as the popular open-domain vocabulary –WordNet. Ontologies like WordNet can be semantically enriched to obtain extensions and enhancements to its lexical database. The proliferation of senses in WordNet is considered as one of its main shortcomings for practical applications. Therefore, the presented methodology depends on the Coarse-Grained word senses. These senses are generated from applying WordNet Fine-Grained word senses to a Merging Sense algorithm. This algorithm merges only semantically similar word senses instead of applying traditional clustering techniques. A performance comparison is illustrated between two different data sources (Web, Corpus) used in the Enrichment process. The results obtained from using Coarse-Grained word senses in both cases yields better precision than Fine-Grained word senses in the Word Sense Disambiguation task.

Keywords: coarse-grained word senses, corpus, ontology, semantic web, word sense disambiguation (WSD), word senses

Categories: M.7