Machine Learning Model for Concrete Strength Prediction

This study assesses machine learning algorithms, including Random Forest, GS-SVR, Multiple Linear Regression, and AdaBoost, in predicting compressive strength of concrete in construction. Utilizing a diverse dataset with feature-enhanced variables, the research aims to provide insights into model performance regarding accuracy, robustness, and efficiency. The dataset is split for unbiased evaluation, and algorithms are selected for their diverse strengths. Random Forest captures complex relationships, GSSVR handles non-linearities, Multiple Linear Regression serves as a benchmark, and AdaBoost explores boosting potential. Performance metrics like MSE, R-squared, Absolute Percentage Error, MAE, and RMS assess model accuracy and efficiency. Results offer information into algorithm strengths and constraints, assisting practitioners in the model selection process. The study contributes to machine learning in construction materials science, providing practical guidance for precise concrete strength predictions and enhancing the quality of infrastructure.