Prediction of Compressive Strength of Geopolymer Concrete by Using Random Forest Algorithm

Abstract

AbstractGeopolymer concrete is a new invention of the concrete industry. It could be the future of all construction fields due to its performance against severe conditions, and strength. It is a perfect alternative to conventional concrete. It is more sustainable, ecological, durable, and economic than conventional concrete. In the present era, machine learning techniques are also the future of all research and development industries. These techniques predict the results based on their previous data. In the construction industry, the find the results or value are very difficult, time consumable, and laborious. These techniques make them very easier to predict the strength of mix design without making samples and destructive tests. The aim of this study is to predict the compressive strength of flyash-based geopolymer concrete by using deep learning and random forest algorithm and comparing them with different errors and coefficient correlation. After the simulation of data, it is proved that the random forest algorithm is the most suitable technique for the prediction of compressive strength. After the developing a model, the various errors were found for accuracy. The mean absolute error, root mean square error, relative absolute error, and root relative squared error are 1.63%, 2.68%, 30.28%, and 37.47%, respectively for the deep learning predicted compressive strength. The errors provide the proof of model accuracy to predict the compressive strength on the basis of ingredients proportions.KeywordsGreen concreteGeopolymer concreteSustainableMachine learningRandom forest algorithm