Prediction of ultimate bearing capacity for rubberized concrete filled steel tube columns based on Tabular Variational Autoencoder method and Stacking ensemble strategy

The application of rubberized concrete filled steel tube (RuCFST) structures can facilitate the recycling of used tires, offering a sustainable solution. In recent years, data-driven machine learning (ML) algorithms have garnered significant attention from researchers in the engineering field. Howev...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Structures (Oxford) Ročník 70; s. 107667
Hlavní autoři: Song, Zongming, Zhang, Chao, Lu, Yiyan
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Ltd 01.12.2024
Témata:
Ensemble learning
Rubberized concrete filled steel tube columns
Stacking
Tabular variational autoencoder
Ultimate bearing capacity
Ultimate bearing capacity
Tabular variational autoencoder
Ensemble learning
Stacking
Rubberized concrete filled steel tube columns
ISSN:2352-0124, 2352-0124
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The application of rubberized concrete filled steel tube (RuCFST) structures can facilitate the recycling of used tires, offering a sustainable solution. In recent years, data-driven machine learning (ML) algorithms have garnered significant attention from researchers in the engineering field. However, the development of ML models for predicting the ultimate bearing capacity of RuCFST columns has been hindered by a lack of sufficient experimental data. To address this limitation, this study develops a novel ML framework aimed at accurately predicting the ultimate bearing capacity of RuCFST columns. This framework integrates an advanced Tabular Variational Autoencoder (TVAE) data augmentation method and a Stacking ensemble strategy. The TVAE method generates reliable synthetic data to enhance the dataset, while the Stacking strategy integrates the strengths of various ML models, including Gradient Boosting Decision Tree, Extreme Gradient Boosting, Light Gradient Boosting Machine, Random Forest, and Ridge, to improve prediction accuracy. The predictive validity of the developed TVAE-Stacking model was assessed against other ensemble methods and commonly used machine learning models. The findings indicated that the TVAE-Stacking model excels in predicting the ultimate bearing capacity of RuCFST columns. This model serves as a valuable reference for applying RuCFST columns in structural engineering.
ISSN:2352-0124
2352-0124
DOI:10.1016/j.istruc.2024.107667