Text-based domain ontology building using Tf-Idf and metric clusters techniques
Yacine Rezgui · The Knowledge Engineering Review · 2007
Abstract The paper describes the methodology used to develop a construction domain ontology, taking into account the wealth of existing semantic resources in the sector ranging from dictionaries to thesauri. Given the characteristics and settings of the construction industry, a modular, architecture-centric approach was adopted to structure and develop the ontology. The paper argues that taxonomies provide an ideal backbone for any ontology project. Therefore, a construction industry standard taxonomy was used to provide the seeds of the ontology, enriched and expanded with additional concepts extracted from large discipline-oriented document bases using information retrieval (IR) techniques.