Predicting the Level of Safety Performance Using an Artificial Neural Network

Title: Predicting the Level of Safety Performance Using an Artificial Neural Network
Creator: Boateng, Emmanuel Bannor; Pillay, Manikam; Davis, Peter
Relation: Advances in Intelligent Systems and Computing
Publisher Link: http://dx.doi.org/10.1007/978-3-030-02053-8_107
Publisher: Springer International
Resource Type: book chapter
Date: 2019
Description: In this study, an artificial neural network model is developed to predict the level of safety performance on construction sites. Adopting an experimental research design, the model employs safety behaviour, near misses, incidents, fatalities, and the safety risk levels as the inputs, while the safety performance level acted as the output. 339 datasets were generated based on expert intuition and professional experiences. A 5-4-1 Multi-Layer Perceptron with back-propagation was sufficient in building the model that has been trained and validated. The results are promising and show good predictive ability. The developed model could help construction and consultancy firms to assess, forecast, and monitor the level of safety performance of construction projects.
Subject: artificial neural network; construction industry; experimental study; safety performance; safety management; prediction
Identifier: http://hdl.handle.net/1959.13/1449982
Identifier: uon:43797
Identifier: ISBN:9783030020538
Language: eng

		Thumbnail	File	Description	Size	Format