A prediction model for safety performance of construction sites using a linear artificial bee colony programming approach. Issue 2 (3rd April 2022)

• Record Type: Journal Article
• Authors: Abbasianjahromi, Hamidreza; Mohammadi Golafshani, Emadaldin; Aghakarimi, Mehdi
• Is Part Of: International journal of occupational safety and ergonomics. Volume 28:Issue 2(2022)
• Page Start: 1265
• Page End: 1280
• Publication Date: 2022-04-03
• Subjects: safety performance -- prediction model -- machine learning -- linear artificial bee colony programing -- construction industry; Human engineering -- Periodicals; Industrial safety -- Periodicals; Human Engineering -- Periodicals; Occupational Health -- Periodicals; Ergonomie -- Périodiques; Sécurité du travail -- Périodiques; Human engineering; Industrial safety; Periodicals; 620.8
• DOI: 10.1080/10803548.2021.1889878
• Languages: English
• ISSNs: 1080-3548

Abstract

Abstract : The reputation of the construction industry as one of the most world's dangerous industries is because of its tremendous accident and fatality rate. In this study, a prediction model for construction projects' safety performance was presented as an alarm system. The proposed model can predict the safety performance at the beginning of the project. To do so, firstly, data preparation and, secondly, developing an automatic regression model were carried out as the two main steps of the study. To collect data, after identifying effective criteria, a questionnaire was developed. Next, an innovative machine learning method, named linear artificial bee colony programming, was used to discover the relationship between the identified criteria and safety performance. The results of one-dimensional and two-dimensional sensitivity analysis show that the four criteria of management commitment to safety, safety training, safety team and safety budget are the most important factors that influence safety performance.