Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Enhancing prediction accuracy of workability and compressive strength of high-performance concrete through extended dataset and improved machine learning models
This study presents a comprehensive study on the prediction of high-performance concrete (HPC) properties using various regression models. The objective was to develop accurate and reliable models that can estimate workability in terms flow and compressive strength of HPC mixtures based on ingredients proportion. The dataset used in this study is collected from literatures study published in reputed journals and extended using the Monte Carlo method to enhance representation and generalization capabilities. Several regression algorithms including Linear Regression, Support Vector Regression, Decision Tree Regression, Random Forest Regression, AdaBoost Regression, Gradient Boosting Regression, XGBoost Regression, and Artificial Neural Network are evaluated for their predictive performance. Model performance is evaluated using metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), R2 Score, and Cross Validations. The results indicate that the Gradient Boosting Regression model outperforms other models in predicting both flow and compressive strength of HPC mixtures. It achieves an R2 of 0.99 for both flow and compressive strength prediction. Feature importance analysis reveals that the water-to-cement (w/c) ratio has the highest influence on compressive strength predictions while water content on flow prediction. The findings of this study highlight the effectiveness of regression models in predicting HPC properties. The developed models can assist in optimizing concrete mixture designs, enhancing quality control processes, and improving the overall performance of HPC in various construction applications.
Enhancing prediction accuracy of workability and compressive strength of high-performance concrete through extended dataset and improved machine learning models
This study presents a comprehensive study on the prediction of high-performance concrete (HPC) properties using various regression models. The objective was to develop accurate and reliable models that can estimate workability in terms flow and compressive strength of HPC mixtures based on ingredients proportion. The dataset used in this study is collected from literatures study published in reputed journals and extended using the Monte Carlo method to enhance representation and generalization capabilities. Several regression algorithms including Linear Regression, Support Vector Regression, Decision Tree Regression, Random Forest Regression, AdaBoost Regression, Gradient Boosting Regression, XGBoost Regression, and Artificial Neural Network are evaluated for their predictive performance. Model performance is evaluated using metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), R2 Score, and Cross Validations. The results indicate that the Gradient Boosting Regression model outperforms other models in predicting both flow and compressive strength of HPC mixtures. It achieves an R2 of 0.99 for both flow and compressive strength prediction. Feature importance analysis reveals that the water-to-cement (w/c) ratio has the highest influence on compressive strength predictions while water content on flow prediction. The findings of this study highlight the effectiveness of regression models in predicting HPC properties. The developed models can assist in optimizing concrete mixture designs, enhancing quality control processes, and improving the overall performance of HPC in various construction applications.
Enhancing prediction accuracy of workability and compressive strength of high-performance concrete through extended dataset and improved machine learning models
Asian J Civ Eng
Tipu, Rupesh Kumar (Autor:in) / Suman (Autor:in) / Batra, Vandna (Autor:in)
Asian Journal of Civil Engineering ; 25 ; 197-218
01.01.2024
22 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
| DOAJ | 2024
Compressive Strength Prediction of Lightweight Concrete: Machine Learning Models
| DOAJ | 2022
Concrete Compressive Strength dataset
| DataCite | 2024