Enhancing risk prediction in high-rise construction: explainable artificial intelligence enabled probabilistic approach

The escalating complexity of high-rise construction projects necessitates advanced risk management strategies to mitigate delays and cost overruns. This paper presents a novel approach integrating XAI with probabilistic and statistical modeling to analyze data collected from high-rise construction projects across the United States. By applying the CWHM and MCMC techniques, we quantify the impacts of regulatory changes, supply chain disruptions, and workforce dynamics on project costs. XAI enhances the transparency and understandability of our predictive models, fostering stakeholder trust and facilitating more informed decision-making. The application of these methodologies predicts potential risks and elucidates the reasoning behind these predictions. Preliminary findings indicate a potential reduction in delay cost overruns through informed, proactive measures. The study contributes to the body of knowledge by offering actionable insights and recommendations for risk mitigation, thereby promoting more resilient and financially sustainable management practices in the high-rise construction sector. The originality of this work lies in linking predictive accuracy with interpretability, enabling stakeholders to make proactive, data-driven decisions that reduce delay-related cost overruns and promote more resilient and sustainable construction management practices.