Estimation of compressive strength of high-strength concrete by random forest and M5P model tree approaches

Abstract

High strength concrete (HSC) define as the concrete that meets a unique mixture of performance uniformity requirements that cannot be reached routinely using conventional constituents and regular mixing, placing, and curing events. The modeling of such type of concrete is very difficult. In this investigation, the performance of the random forest regression and M5P model tree were compared to estimate the 28th day compressive strength of the HSC. Total data set consists of 83 data out of which 70 % of the total dataset used to train the model and residual 30 % used to test the models. The accuracy of the models was depending upon the three performance evaluation parameters which are correlation coefficient (R), root mean square error (RMSE) and maximum absolute error (MAE). The results recommend that random forest regression is more accurate to predict the compressive strength as compare to M5P model tree. Sensitivity analysis indicates that water (W) and Silica fumes (SF) are the most valuable constituents of the HSC and compressive strength mainly depends on these constituents.

Keywords

• high strength concretes
• m5p model tree
• random forest
• silica fume