A machine learning-based life cycle assessment prediction model for the environmental impacts of buildings

With the emerging importance of achieving climate targets and net-zero levels, assessing the environmental sustainability of buildings is of paramount importance. Life Cycle Assessment (LCA) is a popular tool used for such assessment. However, performing LCA for buildings is time-consuming and challenging due to inconsistencies in the databases, software limitations, and data intensiveness, making it a complex tool for decision-making applications. Therefore, this study proposes a methodological framework to develop surrogate LCA models for buildings using modern machine learning (ML) tools such as Multiple Regression and Artificial Neural Networks (ANN). Such a framework improves the application of LCA in environmental decision-making during the planning of building projects by reducing the time, effort, and complexity associated with conducting LCA of buildings. It can be found that the mean absolute percentage error (MAPE) for the tested dataset in the regression-based model is less than 5 percent rendering it a good surrogate model.