A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Study on predicting compressive strength of concrete using supervised machine learning techniques
Applying machine learning methods for predicting the mechanical characteristics of concrete, particularly its compressive strength, is a crucial element in shaping the future of civil engineering. The current study uses gradient boosting (GBM) and light gradient boosting (LGBM) supervised machine learning (ML) approaches in the Python platform to predict the compressive strength of concrete with eight independent variables such as cement content, fly ash, blast furnace slag, water content, super plasticiser, fine aggregate, and coarse aggregate. A total of 1030 datasets are fed to the ML models by adopting 70:30 split ratio for model training and testing purposes. To evaluate the predictive performance of the ML models, various performance metrics, including R2 (coefficient of determination), RMSE (root mean square error), MSE (mean square error), and MAE (mean absolute error), are employed. Based on the results, the LGBM model demonstrates greater efficiency with an R2 value of 0.916 in predicting compressive strength compared to the GBM model (R2 value of 0.898). Moreover, parameter sensitivity analysis shows that cement content (0.191), fine aggregate (0.148), and water content (0.145) play significant role in predicting concrete compressive strength with the used dataset.
Study on predicting compressive strength of concrete using supervised machine learning techniques
Applying machine learning methods for predicting the mechanical characteristics of concrete, particularly its compressive strength, is a crucial element in shaping the future of civil engineering. The current study uses gradient boosting (GBM) and light gradient boosting (LGBM) supervised machine learning (ML) approaches in the Python platform to predict the compressive strength of concrete with eight independent variables such as cement content, fly ash, blast furnace slag, water content, super plasticiser, fine aggregate, and coarse aggregate. A total of 1030 datasets are fed to the ML models by adopting 70:30 split ratio for model training and testing purposes. To evaluate the predictive performance of the ML models, various performance metrics, including R2 (coefficient of determination), RMSE (root mean square error), MSE (mean square error), and MAE (mean absolute error), are employed. Based on the results, the LGBM model demonstrates greater efficiency with an R2 value of 0.916 in predicting compressive strength compared to the GBM model (R2 value of 0.898). Moreover, parameter sensitivity analysis shows that cement content (0.191), fine aggregate (0.148), and water content (0.145) play significant role in predicting concrete compressive strength with the used dataset.
Study on predicting compressive strength of concrete using supervised machine learning techniques
Asian J Civ Eng
Varma, B. Vamsi (author) / Prasad, E. V. (author) / Singha, Sudhakar (author)
Asian Journal of Civil Engineering ; 24 ; 2549-2560
2023-11-01
12 pages
Article (Journal)
Electronic Resource
English
Advanced Machine Learning Techniques for Predicting Concrete Compressive Strength
| DOAJ | 2025
Predicting Confined Compressive Strength of Concrete Using Machine Learning Approach
| Springer Verlag | 2023
Predicting compressive strength of geopolymer concrete using NDT techniques
| Online Contents | 2018
Predicting compressive strength of geopolymer concrete using NDT techniques
| Springer Verlag | 2018