Research Information System
7th International Project and Construction Management Conference, IPCMC 2022, İstanbul, Turkey, 20 - 22 October 2022, pp.1376-1386
Construction safety possesses a considerable challenge to the construction industry. Factual risk assessment (RA) is vital for the effective identification of construction hazards and controlling the associated risks. In this regard, a wide range of machine learning (ML) methods are used over the construction safety records, for achieving a better representative RA. This study first depicts the accident records in form of graph-structured datasets and then utilizes graph representation learning (GRL) methods for obtaining the node embeddings and optimizing a neighborhoodpreserving. To this end, a technique is used to learn the hazard by embedding an unbiased random walk in the neighborhood space of construction accidents. As a result, the developed system is beneficial for the construction site managers to receive recommendations of possible site safety hazards based on the existing hazard scenario.