Predicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretation

Abstract The accurate determination of pile-bearing capacity is crucial in construction projects to ensure the stability and safety of structures built on foundation piles. Nevertheless, the conventional estimation methods used for this purpose tend to be resource-intensive and time-consuming. Machi...

Full description

Saved in:
Bibliographic Details
Main Authors: Adil Khan, Majid Khan, Waseem Akhtar Khan, Muhammad Ali Afridi, Khawaja Atif Naseem, Ayesha Noreen
Format: Article
Language:English
Published: Springer 2025-03-01
Series:Discover Civil Engineering
Subjects:
Online Access:https://doi.org/10.1007/s44290-025-00215-x
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850208558979743744
author Adil Khan
Majid Khan
Waseem Akhtar Khan
Muhammad Ali Afridi
Khawaja Atif Naseem
Ayesha Noreen
author_facet Adil Khan
Majid Khan
Waseem Akhtar Khan
Muhammad Ali Afridi
Khawaja Atif Naseem
Ayesha Noreen
author_sort Adil Khan
collection DOAJ
description Abstract The accurate determination of pile-bearing capacity is crucial in construction projects to ensure the stability and safety of structures built on foundation piles. Nevertheless, the conventional estimation methods used for this purpose tend to be resource-intensive and time-consuming. Machine learning (ML) methods offer a promising alternative to traditional modeling techniques for assessing pile-bearing capacity, providing a more robust and efficient approach to estimating pile-bearing capacity. For this purpose, an advanced ML technique, gene expression programming (GEP), was utilized to predict pile-bearing capacity. GEP is a computational technique that mimics biological gene expression processes to evolve computer programs or models capable of solving complex problems through the iterative generation, selection, and recombination of code segments. A dataset of 472 reinforced concrete piles obtained from literature, was employed for training, and validating the model. The ten most optimal parameters were selected as inputs. To ensure robustness and accurate evaluation, the collected dataset was partitioned into three distinct subsets: the training set (70%), the testing set (15%), and the validation set (15%). In addition to external validation assessment, eight statistical indicators were used to assess the performance and validity of the developed model. The developed GEP model exhibited exceptional performance in estimating pile-bearing capacity, demonstrating a high correlation coefficient value of 0.963 during the training phase and a value of 0.962 during the validation and testing phases. Moreover, a simple empirical formulation has been developed based on GEP to estimate the pile-bearing capacity. The SHapley Additive exPlanation analysis revealed that the pile tip elevation exhibited the highest contribution in estimating the pile-bearing capacity among the various factors considered. In summary, this research presents the application of the GEP approach in forecasting the pile-bearing capacity, offering engineers and practitioners a valuable tool for optimizing foundation design and ensuring the stability and safety of structures.
format Article
id doaj-art-773bfb20bb2d447fbb878d75a8122e1e
institution OA Journals
issn 2948-1546
language English
publishDate 2025-03-01
publisher Springer
record_format Article
series Discover Civil Engineering
spelling doaj-art-773bfb20bb2d447fbb878d75a8122e1e2025-08-20T02:10:13ZengSpringerDiscover Civil Engineering2948-15462025-03-012113010.1007/s44290-025-00215-xPredicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretationAdil Khan0Majid Khan1Waseem Akhtar Khan2Muhammad Ali Afridi3Khawaja Atif Naseem4Ayesha Noreen5Department of Advanced Civil and Structural Engineering, University of BradfordDepartment of Civil Engineering, Southern Illinois University EdwardsvilleDepartment of Civil Engineering, University of Louisiana at LafayetteDepartment of Civil Engineering, University of Louisiana at LafayetteDepartment of Civil Engineering, University of Louisiana at LafayetteDepartment of Civil Engineering, University of Engineering and TechnologyAbstract The accurate determination of pile-bearing capacity is crucial in construction projects to ensure the stability and safety of structures built on foundation piles. Nevertheless, the conventional estimation methods used for this purpose tend to be resource-intensive and time-consuming. Machine learning (ML) methods offer a promising alternative to traditional modeling techniques for assessing pile-bearing capacity, providing a more robust and efficient approach to estimating pile-bearing capacity. For this purpose, an advanced ML technique, gene expression programming (GEP), was utilized to predict pile-bearing capacity. GEP is a computational technique that mimics biological gene expression processes to evolve computer programs or models capable of solving complex problems through the iterative generation, selection, and recombination of code segments. A dataset of 472 reinforced concrete piles obtained from literature, was employed for training, and validating the model. The ten most optimal parameters were selected as inputs. To ensure robustness and accurate evaluation, the collected dataset was partitioned into three distinct subsets: the training set (70%), the testing set (15%), and the validation set (15%). In addition to external validation assessment, eight statistical indicators were used to assess the performance and validity of the developed model. The developed GEP model exhibited exceptional performance in estimating pile-bearing capacity, demonstrating a high correlation coefficient value of 0.963 during the training phase and a value of 0.962 during the validation and testing phases. Moreover, a simple empirical formulation has been developed based on GEP to estimate the pile-bearing capacity. The SHapley Additive exPlanation analysis revealed that the pile tip elevation exhibited the highest contribution in estimating the pile-bearing capacity among the various factors considered. In summary, this research presents the application of the GEP approach in forecasting the pile-bearing capacity, offering engineers and practitioners a valuable tool for optimizing foundation design and ensuring the stability and safety of structures.https://doi.org/10.1007/s44290-025-00215-xPile bearing capacityFoundation designGene expression programmingMachine learningShapley Additive exPlanation
spellingShingle Adil Khan
Majid Khan
Waseem Akhtar Khan
Muhammad Ali Afridi
Khawaja Atif Naseem
Ayesha Noreen
Predicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretation
Discover Civil Engineering
Pile bearing capacity
Foundation design
Gene expression programming
Machine learning
Shapley Additive exPlanation
title Predicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretation
title_full Predicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretation
title_fullStr Predicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretation
title_full_unstemmed Predicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretation
title_short Predicting pile bearing capacity using gene expression programming with SHapley Additive exPlanation interpretation
title_sort predicting pile bearing capacity using gene expression programming with shapley additive explanation interpretation
topic Pile bearing capacity
Foundation design
Gene expression programming
Machine learning
Shapley Additive exPlanation
url https://doi.org/10.1007/s44290-025-00215-x
work_keys_str_mv AT adilkhan predictingpilebearingcapacityusinggeneexpressionprogrammingwithshapleyadditiveexplanationinterpretation
AT majidkhan predictingpilebearingcapacityusinggeneexpressionprogrammingwithshapleyadditiveexplanationinterpretation
AT waseemakhtarkhan predictingpilebearingcapacityusinggeneexpressionprogrammingwithshapleyadditiveexplanationinterpretation
AT muhammadaliafridi predictingpilebearingcapacityusinggeneexpressionprogrammingwithshapleyadditiveexplanationinterpretation
AT khawajaatifnaseem predictingpilebearingcapacityusinggeneexpressionprogrammingwithshapleyadditiveexplanationinterpretation
AT ayeshanoreen predictingpilebearingcapacityusinggeneexpressionprogrammingwithshapleyadditiveexplanationinterpretation