Assessing the Influence of Risk Maturity Level on Construction Delay Factors: A Machine Learning-Based Model

Construction project delays remain a persistent challenge in the construction industry, yet existing studies have generally examined delay risk prioritisation and organisational risk management maturity as separate research domains. This study addresses this gap by investigating how organisational risk management maturity influences the evaluation and prioritisation of construction delay risks in the Australian construction context. A structured survey instrument based on the Generic Risk Maturity Model was used to assess maturity across five dimensions: management commitment and leadership, organisational risk culture, risk identification, risk analysis, and standardised risk management processes. Expert judgements were used to derive impact severity, probability of occurrence, and composite criticality scores for 22 construction delay risk factors. Extreme Gradient Boosting (XGBoost) models with SHapley Additive exPlanations (SHAP) were then applied to explain how maturity dimensions influence risk evaluations. The results show that maturity is associated with changes in both the level and pattern of delay risk prioritisation. Owner’s late decisions remained the top-ranked composite risk across all maturity levels, while other risks exhibited maturity-related shifts, with the composite risk weight of poor project cost estimation declining from 0.079 at maturity Level 1 to 0.041 at maturity Levels 3–4, and rainfall and weather conditions becoming more prominent at higher maturity levels. The proposed XGBoost–SHAP framework provides a transparent, data-driven diagnostic tool for linking organisational capability with delay risk prioritisation and supporting more consistent and evidence-based delay risk management. The results further indicate that weaker organisational risk culture and lower process standardisation are associated with greater perceived exposure to planning-, coordination-, and decision-related delay risks. Within the sampled Australian construction organisations, the XGBoost–SHAP framework provides a transparent analytical approach for examining associations between organisational capability and delay risk prioritisation.