Arabic Ontology Generation By Mapping Between Arabic Wordnet With Arabic Encyclopedia

Abstract

There is great interest in the ontology as a tool for knowledge representation. Today, ontology learning, is considered an important research. That is used in many applications such as: semantic search (Benassi et al., 2004) and automatic summarization (Carenini et al., 2008) and others. There were many researches, projects, models, and tools for this purpose. However, generating an Arab ontology still requires significant research efforts. One of the most widely used Arabic ontologies is the Arabic Word Net (AWN), which was produced by translating the English Word Net into Arabic, and it is used as a lexical ontology in current researches, because of its free availability and coverage. Despite its widespread, it lacks some useful information for Arabic language processing.

Therefore, in this research, a framework for learning Arabic ontology was developed that takes into account the challenges of the Arabic language and the use of Arabic Knowledge sources, and an algorithm was adopted to learn ontology using the mapping technique between the AWN and the features of the Arabic Encyclopedia articles, which is a rich Arabic source of information in all fields of human knowledge. Accordingly, an Arabic ontology was created that combined the comprehensiveness of the AWN with the quality of the Arabic encyclopedia concepts, terminology and Arabic relations. The results showed that a good number of Arabic concepts and terms and semantic relations were added to AWN with good Precision, and therefore the generated ontology can be considered a core Arabic ontology, that can be reused in Arabic semantic applications, and completed to obtain an effective Arabic ontology.