CAOGen: An Automatic Ontology Constructor Based on Data Mining Techniques

CAOGen: An Automatic Ontology Constructor Based on Data Mining Techniques

연구 분야: Databases

학회: International Conference on e-Infrastructure and e-Services for Developing Countries

The construction of ontologies is a classification process in which concepts in a the domain and the relationships between these concepts need to be identified. The classification of concepts to construct an ontology is a difficult problem. The holistic objective is to propose a system that is capable of automatically classifying concepts of a company to construct ontologies. To achieve this, researchers proposed some solutions among which Norms2Onto, Text2onto and APOET (Automatic Product Ontology Extraction from Textual) but majority of them are either domain specific or construct ontologies with generic data which makes the resulting ontology not precise. This paper proposes a system named CAOGen (Company Automatic ontology Generator) that applies CRISPDM (Cross Industry Standard Process for Data Mining) methodology to construct ontologies by using data mining techniques to automatically classify concepts of a specific company to produce its ontology while using wordnet and wikipeadia to augment them. The validation of this work is done through the construction of an ontology of catalogue of service for an Enterprise in Cameroon named yowyob (yowyob.com). After evaluation of the ontology, the system had an accuracy of 0.994, a precision of 0.992, a recall of 1.0 and a F1 score of 0.996 for semantic search and an accuracy of 0.888, a precision of 0.990, a recall of 0.849 and a F1 score of 0.914 for recommendation.