Mining Construction Safety Documents for Safety Concept Structure Discovery Using Formal Concept Analysis

Construction safety documents regulate significant safety actions and requirements by which construction workers or employees should abide in order to secure them from occupational hazard events. Therefore, facilitating faster identification of applicable safety requirements from the documents has become an important topic in the construction safety domain. To address the need in this regard, tools and techniques have been developed and research efforts have been made. One of the research efforts is to utilize ontology, a knowledge representation and reasoning approach, as the methodology to achieve the goal of identifying applicable safety requirements. In such ontology-based researches, the development and construction of the safety concept ontology is an essential task as the concepts and the relationships between the concepts both representing key safety knowledge need to be carefully identified. In this paper, the author focuses on the safety documents without predefined concept structure and aims to address the concept and relationship identification difficulties. Specifically, the author leverages the Formal Concept Analysis (FCA) methodology, a theory of data analysis that identifies conceptual structures among data sets, to mine and analyze the documents in order to discover safety concept structure and to further assist the development of the concept ontology for the construction safety documents. The author expects the application of FCA to construction safety domain can eventually benefit the ontology-based approaches for identification of applicable safety requirements.