Support vector machine applied to prediction strength of cement

Abstract

The prediction strength of cement is an important task in civil engineering. In this study, the support vector machine (SVM), a novel type of learning algorithm based on statistical theory, has been used to predict the 28d strength of cement. The seven input variables used for the SVM model for prediction of strength are content of slag, SO3 content, cement fineness, 1d compressive strength and folding strength, 3d compressive strength and folding strength. Comparison between SVM and artificial Neural network (ANN) methods is also presented. The study shows that the SVM methods can achieve better accuracy and generalization than the ANN methods; and SVM has the potential to be a useful and practical tool for prediction strength of cement.