The Variables Reduction for Concrete Strength Prediction Model Using a Rough Sets Method

Abstract:

Based on the test results of 448 sets of concrete mixtures (which have been divided into three groups), a concrete strength prediction model has been established using a stepwise regression method. The kinds and qualities of raw materials, which has a total comes to 17 factors, has been considered as the independent variables. It has been found that the forecast precision is not always improving along with the increase of independent variable number, so one or more factors may be "redundant" and should be reducted. A rough sets (RS) method has been adopted for the variable reduction, with different parameter combinations are tested. The results has shown that in each cases the forecast precision of the model are higher than the Bowromi formula’s. Some regularities for the parameters’ selections are existed, but not all the parameters can be given positively, so a trial process is still necessary.

• Periodical: Advanced Materials Research (Volumes 311-313)
• Pages: 1911-1915
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.311-313.1911
• Online since: August 2011
• Authors: Bin Chen, Xi Jun Liu
• Keywords: Concrete, Prediction, Rough Set (RS), Strength, Variables Reduction

References

[1] A.M. Neville: Properties of Concrete (Longmen, U.K. 1986).

[2] W.P.S. Dias, S.P. Pooliyadda: Construct Build Mater 4(2001), pp.371-379.

[3] Z. Pawlak: International Journal of Computer Information and Science Vol. 5 (1982), pp.341-356.

[4] R. Slowinski: Intelligent Decision Support: Handbook of Applications and Advances of the Rough Sets Theory (Kluwer Academic Publishers. 1992)

[5] Z.Pawlak: Rough Sets-Theoretical Aspects of Reasoning about Data (Kluwer Academic Publishers, Dordrecht. 1991).

[6] Y.Zhao, F.Luo, S.K.M. Wong and Y.Y. Yao: Proceedings.Toronto.Ont. (Springer, Canada 2007).