A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Application of machine learning algorithms to performance prediction of rocking shallow foundations during earthquake loading
https://orcid.org/0000-0003-2129-8451
Abstract The objective of this research is to develop preliminary predictive models for the performance of rocking shallow foundations using machine learning algorithms and supervised learning technique. Data from a rocking foundation database, consisting of dynamic base shaking experiments conducted on centrifuges and shaking tables, have been used for the development of multivariate linear regression (MLR) model using stochastic gradient descent optimization and distance-weighted k-nearest neighbors (k-NN) regression model. Seismic energy dissipation in soil, permanent settlement of foundation, and the maximum acceleration transmitted to the structure are considered as the performance parameters of rocking foundations, while the input features to machine learning algorithms include critical contact area ratio and rocking coefficient of soil-foundation system, and peak ground acceleration and Arias intensity of earthquake ground motion. It is found that both MLR and weighted k-NN models perform satisfactorily in capturing the complex relationships between the performance parameters and input features of rocking foundations, and that both models perform better than simple statistical models. Based on multiple k-fold cross validation tests across all performance parameters and the mean absolute percentage errors, it is found that the weighted k-NN model consistently outperforms the MLR model in terms of accuracy for the problem considered.
Highlights This study applies MLR and weighted k-NN algorithms to develop predictive models for the performance of rocking foundations. Data mined from a rocking foundations database are utilized in this study using supervised learning technique. The performances of both machine learning models are compared using MAPE and multiple k-fold cross validation technique.
Application of machine learning algorithms to performance prediction of rocking shallow foundations during earthquake loading
https://orcid.org/0000-0003-2129-8451
Abstract The objective of this research is to develop preliminary predictive models for the performance of rocking shallow foundations using machine learning algorithms and supervised learning technique. Data from a rocking foundation database, consisting of dynamic base shaking experiments conducted on centrifuges and shaking tables, have been used for the development of multivariate linear regression (MLR) model using stochastic gradient descent optimization and distance-weighted k-nearest neighbors (k-NN) regression model. Seismic energy dissipation in soil, permanent settlement of foundation, and the maximum acceleration transmitted to the structure are considered as the performance parameters of rocking foundations, while the input features to machine learning algorithms include critical contact area ratio and rocking coefficient of soil-foundation system, and peak ground acceleration and Arias intensity of earthquake ground motion. It is found that both MLR and weighted k-NN models perform satisfactorily in capturing the complex relationships between the performance parameters and input features of rocking foundations, and that both models perform better than simple statistical models. Based on multiple k-fold cross validation tests across all performance parameters and the mean absolute percentage errors, it is found that the weighted k-NN model consistently outperforms the MLR model in terms of accuracy for the problem considered.
Highlights This study applies MLR and weighted k-NN algorithms to develop predictive models for the performance of rocking foundations. Data mined from a rocking foundations database are utilized in this study using supervised learning technique. The performances of both machine learning models are compared using MAPE and multiple k-fold cross validation technique.
Application of machine learning algorithms to performance prediction of rocking shallow foundations during earthquake loading
https://orcid.org/0000-0003-2129-8451
Abstract The objective of this research is to develop preliminary predictive models for the performance of rocking shallow foundations using machine learning algorithms and supervised learning technique. Data from a rocking foundation database, consisting of dynamic base shaking experiments conducted on centrifuges and shaking tables, have been used for the development of multivariate linear regression (MLR) model using stochastic gradient descent optimization and distance-weighted k-nearest neighbors (k-NN) regression model. Seismic energy dissipation in soil, permanent settlement of foundation, and the maximum acceleration transmitted to the structure are considered as the performance parameters of rocking foundations, while the input features to machine learning algorithms include critical contact area ratio and rocking coefficient of soil-foundation system, and peak ground acceleration and Arias intensity of earthquake ground motion. It is found that both MLR and weighted k-NN models perform satisfactorily in capturing the complex relationships between the performance parameters and input features of rocking foundations, and that both models perform better than simple statistical models. Based on multiple k-fold cross validation tests across all performance parameters and the mean absolute percentage errors, it is found that the weighted k-NN model consistently outperforms the MLR model in terms of accuracy for the problem considered.
Highlights This study applies MLR and weighted k-NN algorithms to develop predictive models for the performance of rocking foundations. Data mined from a rocking foundations database are utilized in this study using supervised learning technique. The performances of both machine learning models are compared using MAPE and multiple k-fold cross validation technique.
Application of machine learning algorithms to performance prediction of rocking shallow foundations during earthquake loading
Gajan, Sivapalan (author)
2021-09-01
Article (Journal)
Electronic Resource
English
| British Library Conference Proceedings | 2021
| British Library Conference Proceedings | 2023
| DOAJ | 2024