A Correlation Analysis of Construction Site Fall Accidents Based on Text Mining

Construction site fall accidents are a high-frequency accident type in the construction industry and have received extensive attention from accident causal factor analysis and risk management research, but evaluating the relationship between accident causal factors and unstructured texts remains an area in urgent need of further study. In this paper, an analysis method based on text mining was chosen to analyze and process the collected data of 557 investigation reports of construction site fall accidents in China from 2013 to 2019. First, the accident reports were preprocessed to identify six types and 28 causal factors of fall accidents; subsequently, the 28 causal factors were classified into critical causal factors, subcritical causal factors and general causal factors according to their document frequency. Then, the Apriori algorithm was used to analyze the correlation of construction site fall accidents. Finally, strong association rules were obtained between the accident causal factors and between the causal factors and the types of construction site fall accidents. The results showed that 1) insufficient safety technology training and untimely elimination of hidden danger in safe production were the most frequent accident causal factors in fall accident reports. 2) There were different degrees of strong and weak correlations among the causal factors of construction site fall accidents, among which the higher the importance was, the stronger the correlation. 3) There were strong potential laws between the causal factors and the types of fall accidents, and the combination of some causal factors was most likely to lead to the occurrence of the corresponding accident types. This study scientifically and logically elucidated the inherent risk factors for fall accidents, which provides a theoretical basis for preventing fall accidents in construction projects.