A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Predicting compressive strength of concrete with fly ash and admixture using XGBoost: a comparative study of machine learning algorithms
The prediction of compressive strength in concrete plays a crucial role in the construction industry, as it helps ensure the structural integrity of buildings and infrastructure. This study examines the utilization of fly ash and admixture combinations as input factors for predicting compressive strength using four machine learning algorithms: random forest (RF), support vector machine (SVM), artificial neural network (ANN), and XGBoost, which are utilized and evaluated based on a dataset that is divided into 70% for training and 30% for testing. The performance of each model is assessed by using performance indicators such as mean absolute error, mean squared error, and R-squared. Among all machine learning models, the most effective algorithm for predicting the compressive strength of concrete is XGBoost, which performs better than the other models with an R-squared value of 0.9965 and an RMSE value of 0.9605. Its outstanding performance was aided by its capacity for handling complex relationships between input variables, feature selection abilities, and strong regularization strategies. In terms of accuracy and predictive power, XGBoost exceeded RF, SVM, and ANN, making it a viable technique for predicting concrete's compressive strength. This work advances predictive modelling in concrete engineering, enabling a more accurate compressive strength forecast and eventually enhancing the performance and durability of concrete buildings.
Predicting compressive strength of concrete with fly ash and admixture using XGBoost: a comparative study of machine learning algorithms
The prediction of compressive strength in concrete plays a crucial role in the construction industry, as it helps ensure the structural integrity of buildings and infrastructure. This study examines the utilization of fly ash and admixture combinations as input factors for predicting compressive strength using four machine learning algorithms: random forest (RF), support vector machine (SVM), artificial neural network (ANN), and XGBoost, which are utilized and evaluated based on a dataset that is divided into 70% for training and 30% for testing. The performance of each model is assessed by using performance indicators such as mean absolute error, mean squared error, and R-squared. Among all machine learning models, the most effective algorithm for predicting the compressive strength of concrete is XGBoost, which performs better than the other models with an R-squared value of 0.9965 and an RMSE value of 0.9605. Its outstanding performance was aided by its capacity for handling complex relationships between input variables, feature selection abilities, and strong regularization strategies. In terms of accuracy and predictive power, XGBoost exceeded RF, SVM, and ANN, making it a viable technique for predicting concrete's compressive strength. This work advances predictive modelling in concrete engineering, enabling a more accurate compressive strength forecast and eventually enhancing the performance and durability of concrete buildings.
Predicting compressive strength of concrete with fly ash and admixture using XGBoost: a comparative study of machine learning algorithms
Asian J Civ Eng
Gogineni, Abhilash (author) / Panday, Indra Kumar (author) / Kumar, Pramod (author) / Paswan, Rajesh Kr. (author)
Asian Journal of Civil Engineering ; 25 ; 685-698
2024-01-01
14 pages
Article (Journal)
Electronic Resource
English