Short-term imputation of missing sound level data using selected computational intelligence methods

Abstract

The aim of the paper was to impute for the shortterm missing sound level data in the noise monitoring stations by applying the models which describe variability of sound level within the tested period. To build the model, the computational intelligence methods, like neural networks, fuzzy systems, or regression trees can be used. The latter approach was applied to build the models with the aid of Cubist regression tree and Random Forest regression software, using recorded equivalent sound levels.