Predictive Analysis of Ceramic Waste Modified Concrete Properties Using ANN and Linear Regression Algorithm

In this study, concrete modified with ceramic waste was modelled. The ceramic waste percentage ranged from 2.5% to 5% to 10% to 12.5% to 15% to 17.5% to 20%. Modelling was done for the concrete's tensile strength and compressive strength. Regression modelling and artificial neural networks were used as prediction methods for concrete strength. The models developed in this study to predict the mechanical properties of concrete were evaluated using Mean absolute error, coefficient of determination and root mean square error. The R2 value for the ANN model was determined to be 0.97, compared to 0.95 for the linear regression model. For the one-week, two-week, and four-week prediction models, RMSE values were 1.1 MPa, 1.15 MPa, and 1.05 MPa for the ANN model for one-week, two-week and four-week, respectively, while the linear regression model displayed the RMSE values of 1.08 MPa, 1.22 MPa, and 1.25 MPa. The R2 values for ANN and LR models were estimated to be 0.87 and 0.7, respectively, for predicting split tensile strength. This study will conclude that the artificial neural network model has high accuracy. It can be employed in modelling the mechanical properties of ceramic-modified concrete.