Development of the FieldRisk Algorithm for Predicting Construction and Infrastructure Project Execution Risks Using Supply Chain, Technical, and Operational Data Compared with Random Forest, XGBoost, and Logistic Regression
International Journal of Scientific Research in Civil Engineering · 2023
DOI: 10.32628/ijsrce237217 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Scientific Research in Civil Engineering, 2023. Verfasst von Abdulraheem Onimisi Lawal und 1 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Construction and infrastructure projects generate heterogeneous execution data across procurement, engineering, quality, maintenance, scheduling, cost control, and field operations. Conventional risk registers rarely exploit these streams jointly, while many predictive studies focus on a single outcome such as delay or cost overrun. This technical report develops FieldRisk, a domain-informed classification algorithm that integrates supply-chain, technical, and operational indicators and their interactions into a calibrated execution-risk score. The framework combines engineered project-control variables, domain risk indices, interaction terms, and explainability outputs, and it is benchmarked against Random Forest, XGBoost, and Logistic Regression. The supplied demonstration dataset contains 3,000 project-period observations and 32 initial predictors, with a 60/20/20 train-validation-test split. In the supplied illustrative evaluation, FieldRisk achieved accuracy of 0.91, recall of 0.92, F1-score of 0.90, ROC-AUC of 0.95, PR-AUC of 0.92, and a Brier score of 0.08, compared with ROC-AUC values of 0.93, 0.91, and 0.84 for XGBoost, Random Forest, and Logistic Regression, respectively. Schedule variance, supplier lead-time deviation, productivity deterioration, rework, and equipment downtime were the dominant risk drivers. Ablation results indicated that combining all three data domains produced stronger discrimination than any single domain. FieldRisk therefore provides a technically interpretable architecture for early-warning risk intelligence; however, the numerical evaluation remains illustrative until validated against traceable field data from completed and active projects.
Zitieren
APA
Lawal, A. O., & Enyejo, J. O. (2023). Development of the FieldRisk Algorithm for Predicting Construction and Infrastructure Project Execution Risks Using Supply Chain, Technical, and Operational Data Compared with Random Forest, XGBoost, and Logistic Regression. International Journal of Scientific Research in Civil Engineering. https://doi.org/10.32628/ijsrce237217
Vancouver
Lawal AO, Enyejo JO. Development of the FieldRisk Algorithm for Predicting Construction and Infrastructure Project Execution Risks Using Supply Chain, Technical, and Operational Data Compared with Random Forest, XGBoost, and Logistic Regression. International Journal of Scientific Research in Civil Engineering. 2023. doi:10.32628/ijsrce237217
BibTeX
@article{Lawal2023,
author = {Lawal, Abdulraheem Onimisi and Enyejo, Joy Onma},
title = {Development of the FieldRisk Algorithm for Predicting Construction and Infrastructure Project Execution Risks Using Supply Chain, Technical, and Operational Data Compared with Random Forest, XGBoost, and Logistic Regression},
journal = {International Journal of Scientific Research in Civil Engineering},
year = {2023},
doi = {10.32628/ijsrce237217}
}
EUVIMED