Development of a prediction model for the proportion of buildings exposed to construction noise in excess of the construction noise regulation at urban construction sites

Noise in construction projects, one of the major problems at urban construction sites, should be evaluated and managed according to the construction noise regulation. This study developed a prediction model for proportion of buildings exposed to noise in excess of the construction noise regulation (the overexposed building ratio) as the evaluation index, to effectively manage construction noise using elastic net linear regression and polynomial interpolation. Through the conducted case study, the developed prediction model showed better prediction performance (a 35.5% average increase) than the general model. The developed prediction model was applicable to construction sites with no database and was able to predict not only the overexposed building ratio but also the number of dwellers exposed to such noise level. The prediction model can encourage construction companies and dwellers to actively participate in the management of construction noise and can provide acoustic comfort to communities.