Combining Ontology and Natural Language Processing Methods for Prevention of Falls from Height

Abstract

Safety on construction sites is the most important aspect that a company should guarantee to its employees. To reduce the risk, it is necessary to analyze HIgh POtential (HIPO) hazards. In this study, we focus on Fall From Height (FFH) risk which is one of the main causes of worker fatalities. In order to improve the prevention plan, artificial intelligence (AI) can help to determine the causes and the safety actions from FFH historical data. This paper aims to: (i) develop and populate an ontology for FFH with the help of domain experts, and (ii) analyze and extract key information from a HIPO database through Natural Language Processing (NLP). Experimental results are conducted in order to evaluate the proposed approach.