Answers are grounded in this paper and cite their sources.
Accurate prediction of concrete compressive strength is essential for material design, structural safety assessment, and quality control in modern construction engineering. Although machine learning techniques have demonstrated considerable potential for strength prediction tasks, most existing approaches employ fixed learning strategies and treat heterogeneous concrete samples uniformly, limiting their ability to adapt to variations in mixture compositions and curing conditions. Consequently, predictive performance may deteriorate when the relationships between input variables and strength development vary across operating conditions. To address this challenge, a novel Condition-Aware Dual-Adaptive Ensemble (CADAE) framework is proposed for concrete compressive strength prediction. The framework introduces a condition-aware learning mechanism that dynamically adjusts ensemble behavior according to key concrete-related conditions, enabling more flexible modeling of heterogeneous strength-development patterns. In addition, a dual-adaptive fusion strategy is developed to jointly enhance global predictive consistency and local error correction, thereby improving both prediction accuracy and robustness. To comprehensively evaluate the effectiveness of the proposed framework, extensive experiments were conducted, including benchmark comparisons, prediction-behavior analyses, ablation studies, component-contribution investigations, residual-distribution evaluations, and parameter-sensitivity analyses. The proposed CADAE framework consistently outperformed conventional machine learning and ensemble-learning models across multiple evaluation metrics, achieving an R 2 of 0.9564, RMSE of 3.1745 MPa, MAE of 2.2258 MPa, and MAPE of 7.85%. Compared with the strongest baseline model (ExtraTrees, R 2 = 0.9313, RMSE = 3.9865 MPa), CADAE improved R 2 by 0.0251 (2.7% relative improvement) and reduced RMSE by 20.4%. Ablation studies further confirmed the effectiveness of individual modules, while sensitivity analyses demonstrated stable performance under various parameter configurations. Moreover, the adaptive learning behavior exhibited strong consistency with engineering principles related to concrete strength development, providing enhanced practical interpretability. Overall, the proposed framework offers an accurate, robust, and scalable solution for concrete compressive strength prediction and presents a promising paradigm for integrating condition-aware learning and adaptive ensemble modeling in intelligent construction-material applications.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: