Data Mining Mechanical properties of concretes containing calcined clays as supplementary cementitious materials in concrete

Concrete due to its excellent properties, low cost, and wide availability is the most used construction product in the world. As the world population grows, cement production has increased as one of the main materials used in concrete. Cement production emits carbon dioxide which increases environmental pollution. One of the ways to prevent pollution is to use Supplementary cementing materials instead of cement. One of the most important alternative materials used in recent years is LC3 cement. This type of cement reduces the amount of clinker in cement, decreases our need for fossil fuels, and therefore reduces the emission of carbon dioxide. Moreover, one of the most important mechanical properties of concrete is its compressive strength, which its estimation is complex since the number of existing parameters is high. Hence, costly laboratory methods are used which consist of high error. In this paper, machine learning models that work based on decision tree model are used which their performance have been improved by genetic algorithm. LightGBM and XGBoost models got a prediction score of 0.958 in the prediction of the compressive strength of concrete and perform better than decision tree and random forest models with 0.91 and 0.932 prediction scores. Also, the feature importance of each model has been presented and a new data set has been used to evaluate the validation of models.