article · International Journal of Scientific Research in Civil Engineering
Construction organisations typically assess environmental, occupational health and safety, and quality audit findings independently, which hides connections across standards. To address this, a unified model called the Integrated ISO Sustainable Construction Compliance Algorithm was created. The framework merges indicator weighting, graph-based dependency modelling, and explainable gradient boosting. Using a dataset of 6,120 audit findings across ISO 14001, ISO 45001, and ISO 9001 standards, the algorithm mapped overlaps between safety, quality, and environmental issues. The model achieved 96.8% accuracy and 98.7% area under the receiver operating characteristic curve, outperforming conventional machine learning methods. Key factors influencing compliance included delays in corrective actions, recurring critical issues, and links between safety and quality failures. The resulting system provides predictive, data-driven support for managing risks, benchmarking projects, and overseeing regulatory compliance.
Construction projects must balance safety, environmental impact, and build quality to remain compliant and sustainable. Treating these areas as separate audit categories risks missing systemic problems that cause accidents or project delays. By examining audit records as an interconnected system, managers and regulators can predict risks earlier and resolve the root causes of recurring site problems more effectively.
The algorithm could be incorporated into compliance software, risk management tools, or audit platforms used by construction firms and regulatory bodies. The technology is applied and tested on a historical dataset of audit records, demonstrating clear decision-support utility, but it requires integration into live project management systems to reach full commercial deployment.
AI-generated from the published abstract. Always read the original work before citing.
Construction organizations generate substantial audit evidence under ISO 14001, ISO 45001, and ISO 9001, yet environmental, occupational health and safety, and quality findings are frequently evaluated independently. This fragmentation conceals cross-standard dependencies and limits predictive compliance management. This study developed the Integrated ISO Sustainable Construction Compliance Algorithm (IISCCA), a novel framework combining entropy-based indicator weighting, graph-based dependency modeling, and explainable gradient boosting. The dataset comprised 6,120 audit findings, including 2,081 ISO 14001 findings, 2,356 ISO 45001 findings, and 1,683 ISO 9001 findings. Minor, major, and critical nonconformities represented 61.2%, 23.6%, and 15.2% of observations, respectively. Cross-standard analysis produced Jaccard coefficients of 0.64 between safety and quality, 0.58 between environmental and safety, and 0.47 between environmental and quality findings. The Sustainable Construction Compliance Index ranged from 41.8 to 96.4, with a mean of 78.6 ± 11.7. IISCCA was trained using normalized audit indicators, entropy weights, graph centrality measures, recurrence, severity, regulatory significance, defect density, and corrective-action closure time. It achieved 96.8% accuracy, 96.3% precision, 96.7% recall, a 96.5% F1-score, and 98.7% AUROC. IISCCA outperformed XGBoost, artificial neural network, Random Forest, support vector machine, and logistic regression models. SHAP analysis identified corrective-action delay, critical nonconformity recurrence, safety–quality graph centrality, regulatory significance, and defect density as the dominant predictors. The proposed framework provides transparent, robust, and data-driven decision support for project benchmarking, risk prioritization, corrective-action management, regulatory oversight, and sustainable construction improvement.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.32628/ijsrce2594012
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
