Abstract Innovations move more fast in the construction sector, despite challenges including time and money limits. Researchers develop solutions, such as mathematical models and numerical simulations, with the goal of lessening these constraints. Accurately predicting the strength of concrete is essential because modernistic construction techniques have a direct impact on cost effectiveness, structural integrity, and resource efficiency. Conventional methods for estimating strength parameters, which are often constrained by linearity assumptions and small data dimensions, foreshadow regression models and empirical formulas. During recent years, machine learning (ML) models have appeared as a cutting-edge tactic to increase prediction accuracy. ML can handle both complicated algorithms and big datasets. Numerous methods and advancements in this sector are covered in this review article. It provides a comprehensive examination of how machine learning algorithms predict concrete strength parameters. The paper emphasizes how Several machine learning (ML) methodologies, including support vector machines, random forests, gradient boosting techniques, artificial neural networks (ANNs), and others, can be used to model non-linear relationships in complicated datasets. A detailed description of the input parameters frequently employed in these models, including as the w/c, cement content, aggregate size, admixtures, and curing conditions, is provided in parallel with the implications of feature selection and engineering on model performance. Finally, depending on the availability of training data, this review suggests suitable machine learning applications for forecasting concrete strength.
Rao et al. (Fri,) studied this question.