Machine Learning Applied to Acoustic Insulation Analysis in Residential Buildings - Part 2: Vertical partitions

Abstract

The use of Machine Learning (ML) tools has grown a lot in several areas, including in the civil construction space. Part one of this study tried to use ML to estimate results for sound insulation of floor systems and had concluded that ML is not applicable to those situations. Part two tried some adjustments to the ML script and used an on-site vertical partition system test database as input with ML tools used to estimate new results based on the input data. The conclusion of this new study is that the tool is now more applicable to floor systems but it is not the best form to predict the results, given the complexity of sound propagation between rooms, which is not considered in the statistical analysis of the machine.