Abstract
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.
Highlights
With the steady advancement of the national economy and urbanization, the development of the construction industry shows vigorous vitality and rapid expansion (Yiu et al, 2019)
The use of the text mining method to analyze construction site fall accidents encompasses two steps: extraction of accident feature information and correlation analysis of the relevant information pertaining to the accident
The jiebaR package is downloaded, and the worker function is configured with a hidden Markov model (HMM) word segmentation method designed for text segmentation
Summary
With the steady advancement of the national economy and urbanization, the development of the construction industry shows vigorous vitality and rapid expansion (Yiu et al, 2019). Due to the rapid development of engineering construction, the continuous expansion of construction scale and the diversification of structural design have an impact on safe production practice and risk management in the process of project implementation (Choe and Leite, 2016; Hwang et al, 2018; Nawaz et al, 2019). The security problems involved in the process of infrastructure construction adversely affect the main stakeholders of a project, so improvement measures are put forward from the three levels of the government, enterprises and individuals, with the aim of identifying and Correlation Analysis of Falling Accidents
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have