Bayesian learning approach to foresee compressive strength of concrete

One of the key material characteristics that affects the construction of concrete structures is concrete's compressive strength. As long as concrete remains the most popular building material, the occurrence of engineering catastrophes brought on by incorrect mix proportions will remain to be very common. The traditional systems of finding concrete compressive strength are very time consuming resulting the conflict between accurate and timely prediction. This paper deals with Bayesian learning approach including modelling of Ridge, Bayesian and Kernel Ridge Regression (KRR) for the 28-day compressive strength prediction of several concrete mixtures. The analysis of performance evaluates based on R²score, RMSE, MSE, MAE, MAPE, and the corresponding graphs. The R²score of Bayesian regression, Ridge regression, and KRR are 0.66, 0.661, and 0.6167 respectively. Although, this performance index shown that Bayesian regression and Ridge regression outperform KRR, the Bayesian regression model demonstrates a slightly improved performance compared to all proposed methods.