Employing machine learning in water infrastructure management: predicting pipeline failures for improved maintenance and sustainable operations DOI Creative Commons

Yasin Asadi

Industrial Artificial Intelligence, Год журнала: 2024, Номер 2(1)

Опубликована: Ноя. 14, 2024

This study explores techniques for managing class imbalance in predictive modeling to forecast water pipe failures using XGBoost and logistic regression. Given the significant challenges posed by pipeline failures—such as service disruptions, costly repairs, environmental hazards—there is a pressing need effective models. Using dataset from 2015 2022 that includes features like age, material, diameter, maintenance history, applies methods such random oversampling undersampling improve model performance. Results show outperforms regression recall (0.795 vs. 0.683), critical metric infrastructure. Although has slightly better precision (0.695), demonstrates superior overall performance with higher Matthews correlation coefficient (MCC) F1 score, effectively balancing recall. research essential it addresses robust models anticipate mitigate failures. By offering comprehensive framework large-scale datasets showcasing how accurate predictions can reduce costs wastage, this contributes more efficient sustainable infrastructure management.

Язык: Английский

Barriers and Drivers to Implement Alternative Water Use in the Chemical Industry: A Stakeholder Perspective DOI
Haniye Safarpour, Miriam Tariq, Lynn E. Katz

и другие.

Journal of Cleaner Production, Год журнала: 2025, Номер unknown, С. 145582 - 145582

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0

Cost analysis and site selection for reclaimed water injection to enhance coastal aquifer sustainability DOI
Selim Doğan, Shapour Azarm

Journal of Water Process Engineering, Год журнала: 2024, Номер 68, С. 106551 - 106551

Опубликована: Ноя. 15, 2024

Язык: Английский

Процитировано

1

Employing machine learning in water infrastructure management: predicting pipeline failures for improved maintenance and sustainable operations DOI Creative Commons

Yasin Asadi

Industrial Artificial Intelligence, Год журнала: 2024, Номер 2(1)

Опубликована: Ноя. 14, 2024

This study explores techniques for managing class imbalance in predictive modeling to forecast water pipe failures using XGBoost and logistic regression. Given the significant challenges posed by pipeline failures—such as service disruptions, costly repairs, environmental hazards—there is a pressing need effective models. Using dataset from 2015 2022 that includes features like age, material, diameter, maintenance history, applies methods such random oversampling undersampling improve model performance. Results show outperforms regression recall (0.795 vs. 0.683), critical metric infrastructure. Although has slightly better precision (0.695), demonstrates superior overall performance with higher Matthews correlation coefficient (MCC) F1 score, effectively balancing recall. research essential it addresses robust models anticipate mitigate failures. By offering comprehensive framework large-scale datasets showcasing how accurate predictions can reduce costs wastage, this contributes more efficient sustainable infrastructure management.

Язык: Английский

Процитировано

0