Author Identifier
Sanjay Kumar Shukla: https://orcid.org/0000-0002-4685-5560
Document Type
Journal Article
Publication Title
Cogent Engineering
Volume
12
Issue
1
Publisher
Taylor & Francis
School
School of Engineering
Publication Unique Identifier
10.1080/23311916.2025.2467144
Abstract
Seismic analysis often involves significant uncertainty and requires detailed observations. The traditional approaches are constrained by unclear mechanisms and imprecise models to predict the stability of geostructures. The research gap between the accuracy of observed and predicted values can be bridged by employing artificial intelligence-based machine learning (ML) models. The seismic displacement of the nailed soil wall obtained from experimental studies were assessed using suitable ML approaches. Laboratory studies revealed that the critical acceleration was increased by 32% on the inclusion of nails of reinforcement length to excavation height ratio (L/H) to 0.6, and by 17% when the (L/H) was increased from 0.6 to 0.8. The research was focussed on the incorporation of ML techniques for effective data processing and modeling algorithms to extract data. The influence of the design parameters on the seismic stability of nailed soil excavation is investigated using multivariate regression analysis. The ML algorithms, RBF kernel and random Fourier features consistently delivered strong performance pairing with linear regression algorithm. The optimal prediction model indicated 11.34% variation from the target variables with a positive R-squared value. A comparative analysis demonstrated the applicability and generalizability of ML modeling for outcome prediction, benefiting similar studies.
DOI
10.1080/23311916.2025.2467144
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Comments
Muthukumar, S., Sathyan, D., & Shukla, S. K. (2025). Machine learning-based seismic response forecasting using feature mapping algorithms and scientometric analysis of nailed vertical excavation in a soil mass. Cogent Engineering, 12(1), 2467144. https://doi.org/10.1080/23311916.2025.2467144