Ontology Development for Green Building by Using A Semi-automatic Method

Green building has been deemed an important endeavor to promote sustainable building development. However, knowledge from different standards, different companies, and different software in the green building domain is difficult to share and reuse since different terminologies, measurement indicators, and criteria are adopted. Therefore, there is a need to create a consistent knowledge representation model in the green building domain. This study proposes a green building ontology (GB-Onto) which is an abstract conceptualization of the knowledge in the green building domain. To build the ontology more effectively, this study adopts the ontology learning method which is based on NLP and machine learning techniques. An improved TF-IDF method is introduced to extract concepts in the green building domain. Concept inclusion and semantic networks method are integrated to extract taxonomic relations. The associate rule method is used for extracting non-taxonomic relations. Finally, all these methods are implemented by adopting software and Python programming. The GB-Onto is evaluated through consistency checking and criteria-based evaluation. The GB-Onto fills the knowledge gap by providing a formal and shared vocabulary for the green building domain which promotes knowledge reuse and sharing among different stakeholders.