Predicting Compressive Strength and Density of Ecuadorian Natural Mordenite-Based Geopolymer Using Machine Learning Algorithms

Nowadays, in the construction industry, it is crucial to utilize machine learning (ML) to predict the properties of construction materials, minimize potential errors, and improve the desired features. This study focused on estimating the compressive strength (CS) and density of mordenite-based geopolymers using ML models. Optimization techniques are used to improve the performance of the models. Extra Tree using Bayesian optimization is the best model for CS. The mentioned ML model achieved a determination coefficient of 0.8852 with prediction errors MAE=0.8649, MSE=1.3932, RMSE=1.1803, RMSLE=0.3884, and MAPE=6.1239. Similarly, Extra Tree using Bayesian tuning is the best model for density, yielding a determination coefficient of 0.6791 with prediction errors MAE=0.0404, MSE=0.0024, RMSE=0.0494, RMSLE=0.0178, and MAPE=0.0228.