Abstract
Predicting the compressive strength and CO2 uptake of CO2-cured cementitious materials is challenging because these properties are influenced by complex interactions among mixture composition, curing conditions, and carbonation behavior. Accurate prediction and optimization of CO2 curing efficiency are therefore essential for developing sustainable low-carbon cementitious materials. This study developed a Least Squares Boosting (LSBoost)-based machine-learning framework for predicting and optimizing compressive strength and CO2 uptake using a literature-derived database consisting of 423 compressive strength datasets and 131 CO2 uptake datasets. The input parameters incorporated binder chemistry through the Steinour formula, mixture proportions, pre-conditioning conditions, and CO2 curing variables. Hyperparameter tuning identified an optimal LSBoost configuration consisting of 10 splits, a minimum leaf size of 3, a minimum parent size of 10, 500 learning cycles, and a learning rate of 0.5. The developed models demonstrated strong predictive performance, with testing correlation coefficients exceeding 0.95 for both compressive strength and CO2 uptake predictions. The compressive strength model achieved a testing mean absolute percentage error of approximately 10.4%, while the CO2 uptake model yielded a testing RMSE of 3.76 despite the presence of zero-uptake datasets. Feature importance analysis revealed that compressive strength was mainly governed by the Steinour formula, water-to-binder ratio, curing duration, and curing age, whereas CO2 uptake was strongly influenced by CO2 concentration, relative humidity, and exposure duration during CO2 curing. Monte Carlo-simulation-based optimization further suggested that CO2 concentrations above approximately 15%, relative humidity above 60%, and curing durations longer than 48 h promoted higher predicted CO2 uptake under the investigated conditions.
Recommended Citation
S. Han et al., "Machine Learning-Based Prediction Of Compressive Strength And CO2 Uptake In CO2-Cured Cementitious Materials," International Journal of Concrete Structures and Materials, vol. 20, no. 1, article no. 86, SpringerOpen, Dec 2026.
The definitive version is available at https://doi.org/10.1186/s40069-026-00956-8
Department(s)
Civil, Architectural and Environmental Engineering
Publication Status
Open Access
Keywords and Phrases
CO2 curing; CO2 sequestration; CO2 uptake; Least squares boosting model; Optimization
International Standard Serial Number (ISSN)
2234-1315; 1976-0485
Document Type
Article - Journal
Document Version
Final Version
File Type
text
Language(s)
English
Rights
© 2026 The Authors, All rights reserved.
Creative Commons Licensing
This work is licensed under a Creative Commons Attribution 4.0 License.
Publication Date
01 Dec 2026
Comments
National Research Foundation of Korea, Grant RS-2024-00459490