Risks assessment in the construction of infrastructure projects using artificial neural networks

Abstract

Infrastructure Project (IP) construction in Egypt is saddled with copious risks emanating from the manque of risk knowledge, driving these projects to enigmatic failure concerning construction cost. The paper’s novel contribution is to acquaint with an Infrastructure Neural Risk Model (INRM) approach that colossally spotlights the substantial risk factors confronting Egypt’s IP construction apropos to Probability of Occurrence (POO) and Impact on Cost (IOC). Hundred and fifty-seven risk factors were compiled from the literature under three risk categories and reduced to 10, portraying the critical risk influencing the IP construction costs inconsolably. The paper’s main merits include identifying the puritanical risk factors impacting the IP construction costs and bestowing contractors with a paradigm for anticipating risk factors POO and their respective IOC among the IP. Five chronological steps constitute the developed approach inaugurating with (1) conducting a thorough analysis of risk management studies in the IP, (2) adopting a Risk Breakdown Structure (RBS), (3) implementing a checklist analysis to recall the substantial IP risks, (4) constituting a questionnaire survey to interrogate the forthcoming risk factors inferred from comprehensive prior research, and (5) developing INRM paradigm to anticipate the prospective risk factors post-mitigation impact on the construction cost. According to the findings, the coefficient of determination (R2) of the developed INRM unearthed the best INRM outcomes, and the optimal architecture in training and test datasets is 0.872 and 0.777, respectively.

Keywords:

• Infrastructure projects
• neural network
• risk management
• risk assessment

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding details

No grant was given to this research by funding organizations in the public, private, or not-for-profit sectors.

Data availability statement

On request, the corresponding author will provide all or some of the data, models, or algorithms that support the study’s findings.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.