MACHINE LEARNING-BASED MODEL FOR PREDICTING CONCRETE COMPRESSIVE STRENGTH

This study aims at applying a machine learning-based model to establish the relationshipbetween different input variables 28-day compressive strength of normal and High-PerformanceConcrete (HPC). An Artificial Neural Network (ANN) was trained, validated, tested using acomprehensive database consisted 361 records gathered from previously circulated source. Variousmodels with learning algorithms neuron numbers in hidden layer were examined attain thebest performance model. The examination results revealed that ANN “trainlm” learningalgorithm delivered best prediction outcomes overall coefficient determination (R2) 0.9277.The influence parameters on output also by performing sensitivity analysis. Itwas observed concrete 28 days more responsive changes thecement parameter (CM) amount water (WT). In contrast, strengthwas found less sensitive variation fly ash (FL) parameter.