Prediction of the compressive strength of normal and high-performance concretes using M5P model tree algorithm

Abstract

Compressive strength of concrete is one the parameters required in many design codes. A reliable prediction of it can save in time and cost by quickly generating the needed design data. In addition, it can reduce the material waste by reducing the number of trial mixes. In this study, M5P model tree algorithm was used to predict the compressive strength of normal concrete (NC) and high performance concrete (HPC). Compared to other soft computing methods, model trees are able to offer two main advantages: (a) they are able to provide mathematical equations and offer more insight into the obtained equations and (b) they are more convenient to develop and implement. To develop the model tree, a total of 1912 distinctive data records were collected from internationally published literature. Overall, the results show that M5P model tree can be a better alternative approach for prediction of the compressive strength of NC and HPC using the amount of constituents of concrete as input parameters.