A Hybrid Yoruba Noun Ontology
The primary purpose of ontological annotation in Natural Language Processing (NLP) is to make explicit the content of concepts in a domain, using a set of arbitrary tags or labels which must conform to an agreed standard. This method has contributed to the development of many European languages, but it has not been so much employed in African languages. This paper therefore develops a model to prepare the implicit knowledge of Yorùbá noun as found in the literatures into their component features so as to make it both human and machine readable. The ontology development process proposed by (Schultz, 1997) which is based on the activities identified in the IEEE standard for software development is the method of ontological annotations adopted to design the hybrid model for Yoruba noun. Bamgbose (1990)’s “Fonó̩ló̩jì àti Gírámà Yorùbá” , Awobuluyi (1978)’s Essentials of Yorùbá Grammar and Awobuluyi (1978)’s È̩kó̩ Gírámà Èdè Yorùbá were purposively selected and the implicit knowledge of Yoruba noun in these literatures were extracted as the data for implementing the model. The informally perceived domain knowledge of Yorùbá nouns were extracted randomly, using intermediate representations based on tabular and entity-relation notations. Protégé 4.5, a semantic web editing tool was used to implement the hybrid model for Yorùbá Nouns. The model at the final edge specifies the semantic load and properties of each of the nouns captured according to the Yoruba Language Scholars’ perspectives and classifications. The definitions in the annotation serve as backbones for machine learning, web searching and artificial intelligence agents and other (NLP) systems. The model through the annotation schema and label serves as repositories of data and make the content available for machine manipulation and semantic web reasoning. This paper recommends that more on ontological annotations for Yorùbá grammar concepts are needed to foster Yorùbá language engineering.
Keywords: Yorùbá noun, Ontological annotation, Semantic web, Natural language processing