Random Forest-Based Ensemble Estimator for Concrete Compressive Strength Prediction via AdaBoost Method
As one of the most important building materials, the quality of concrete directly affects the safety of buildings. Hence, it is an important and hot issue to predict the compressive strength of concrete with highly accuracy. Most of existing methods heavily depend on building a single model to predict the compressive strength of concrete. However, the proposed single models are not one-size-fits-all, different model have different limitations, making them a better or worse fit for different situation. To address this issue, this paper proposes a novel predict model for concrete compressive strength based on ensemble learning method. In detail, we build our ensemble framework via using the AdaBoost method, while the random forests methods as weak classifier are integrated to the AdaBoost framework. For dealing with the noisy and missing value problems, a set of statistical methods are employed. Furthermore, we utilize the Pearson correlation coefficient to analysis the relationship between different input materials, which can effectively drop out the irrelevant and redundant features. Experimental results on two industrial data sets show that proposed ensemble estimator can significantly improve the prediction accuracy in comparing with other five state-of-the-art methods.
KeywordsConcrete compressive strength Ensemble learning AdaBoost Random Forests Prediction
This work was supported in part by Chongqing research program of key standard technologies innovation of key industries under grant cstc2017zdcy-zdyfX0076, in part by Chongqing research program of technology innovation and application under grant cstc2018jszx-cyztzxX0025 and cstc2017rgzn-zdyfX0020, in part by Youth Innovation Promotion Association CAS, No. 2017393, in part by the National Natural Science Foundation of China under Grant 61602434.
- 2.Popovics, S.: History of a mathematical model for strength development of Portland cement concrete. Mater. J. 95(5), 593–600 (1998)Google Scholar
- 3.Rougeron, P., Aïtcin, P.C.: Optimization of the composition of a high-performance concrete. Cem. Concr. Aggreg. 16(2), 115–124 (1994)Google Scholar
- 8.Schapire, R.E.: The strength of weak learnability. Mach. Learn. 5(2), 197–227 (1990)Google Scholar
- 9.Freund, Y., Schapire, E.R.: A desicion-theoretic generalization of on-line learning and an application to boosting. Computational Learning Theory (1995)Google Scholar