Exploring sustainable solutions with machine learning algorithms: a focus on construction waste management

The accurate estimation of construction waste (CW) is vital for promoting sustainable construction practices. Traditional estimation approaches rely on manual calculations that consider factors such as floor area, building type, and other relevant parameters. Empirical equations are then employed to estimate the quantity of CW generated, but the accuracy of these strategies is typically low. However, machine learning (ML) algorithms have made significant breakthroughs in estimation accuracy, exceeding 90%. This paper investigates trained ML models that can be employed in CW estimation, using a Tabuk city, Saudi Arabia, as a case study. The models undergo rigorous validation and testing through a variety of evaluation metrics such as residuals, mean square error, root mean square error, and coefficient of determination (R²-score). The R²-score values, ranging from 0.88 to 0.98 for both the training and testing datasets, demonstrate that ML algorithm models can make reliable and accurate predictions of CW quantities. Furthermore, these results indicate that utilizing these algorithms can help in developing efficient techniques for managing CW, leading to reduced environmental, economic, and societal effects. The potential applications for these ML algorithms in the construction sector are extensive and mark a promising step forward in promoting sustainability and implementing effective waste management practices within the industry in Saudi Arabia and beyond.