Deep Learning Projections for High-Performance Concrete Strength Forecasting

RK Tipu, OA Shah, S Vats… - 2024 4th International …, 2024 - ieeexplore.ieee.org
RK Tipu, OA Shah, S Vats, S Arora
2024 4th International Conference on Innovative Practices in …, 2024ieeexplore.ieee.org
This study explores the application of a Multilayer Perceptron (MLP) deep learning model to
predict the compressive strength of High Performance Concrete (HPC). The dataset,
comprising 1030 samples, undergoes meticulous preprocessing, including normalization
within a 0-1 range. The MLP model is trained and evaluated, yielding impressive
performance metrics such as an R 2 score of 0.918201, MSE of 13.76, RMSE of 3.71 MPa,
MAE of 4.79 MPa, and a low MAPE of 0.02. Additionally, a comprehensive 10-fold cross …
This study explores the application of a Multilayer Perceptron (MLP) deep learning model to predict the compressive strength of High Performance Concrete (HPC). The dataset, comprising 1030 samples, undergoes meticulous preprocessing, including normalization within a 0-1 range. The MLP model is trained and evaluated, yielding impressive performance metrics such as an R 2 score of 0.918201, MSE of 13.76, RMSE of 3.71 MPa, MAE of 4.79 MPa, and a low MAPE of 0.02. Additionally, a comprehensive 10-fold cross-validation ensures the model's robustness and generalization ability. Visualizations in actual versus predicted plots and radar plots provide intuitive insights into the model's predictive prowess. Hyperparameter tuning further refines the model. The study establishes the MLP model as a reliable tool for predicting HPC compressive strength, offering significant implications for enhancing the precision and efficiency of concrete engineering practices.
ieeexplore.ieee.org
以上显示的是最相近的搜索结果。 查看全部搜索结果