A comprehensive model for concrete strength prediction using advanced learning techniques
Document Type
Article
Publication Title
Discover Applied Sciences
Abstract
Due to environmental concerns and resource limitations, the construction industry faces increasing pressure to adopt sustainable methods. This study proposes a novel hybrid machine learning framework to predict the power of eco-friendly concrete containing eco-friendly concrete, copper slag and eggshell powder as partial cement replacement. Experimental tests were conducted to evaluate compressed and stress powers during various treatment periods. The main ingredient analysis (PCA) was used to reduce the dimension, while random forest regression (RFR), support vector regression (SVR), and the Convolutional Neural Network (CNN) were applied for the forecast. The proposed hybrid PCA—RFR—SVR—CNN MO Model Dale received the best performance, with an average error of 2.0 MPA and a R2 of 0.95. These results show a significant improvement in individual models, providing a strong and accurate tool to predict solid power and support the development of durable construction materials.
DOI
10.1007/s42452-025-07095-x
Publication Date
6-1-2025
Recommended Citation
Dhengare, Sagar; Waghe, Udaykumar; Yenurkar, Ganesh; and Shyamala, Anjana, "A comprehensive model for concrete strength prediction using advanced learning techniques" (2025). Open Access archive. 13134.
https://impressions.manipal.edu/open-access-archive/13134