Prediction of compressive strength of geopolymer concrete using random forest machine and deep learning

Geopolymer concrete (GPC) is a revolutionary innovation in the concrete industry. Due to its resistance to extreme conditions and tensile strength, it may be the future of all construction disciplines. It is an ideal substitute for ordinary concrete. It is more durable, environmentally friendly, sustainable, and cost-effective than traditional concrete. In the modern era, machine learning techniques are the fortune of all research and development industries. These methods predict the outcomes based on their historical data. Finding results or value in the construction industry is difficult, time-consuming, and laborious. These methods make it much simpler to anticipate the potency of a mixture without taking samples or conducting destructive experiments. This study's objective is to predict the strength of GPC by employing deep learning and the random forest algorithm and comparing their errors and coefficient correlations. After data simulation, it is demonstrated that the random forest algorithm is the optimal method for predicting compressive strength. The mean absolute error, root mean square error, relative absolute error, and root relative square error for random forest techniques are 1.63%, 2.68%, 30.28%, and 37.47%, respectively, whereas they are 3.46%, 5.94%, 64.32%, and 82.9% for deep learning techniques. The random forest technique is more precise than the deep learning method because the random forest predicts data with less error than deep learning. Random forest predicted value correlation coefficient is greater than deep learning predicted value correlation coefficient. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.