Missing Data Reconstruction in Acoustic Level Long Term Monitoring

The acoustic level long term monitoring is a common practice in large urban areas, in which, according to the international regulation, the noise levels must be kept under certain thresholds. Sometimes, computational methods are used to predict acoustic level values in future periods. These models need to be calibrated on a continuous measurements dataset and, if some data are missing, the calibration can fail or can be “biased”. In this paper, the problem of missing data reconstruction is approached by means of two techniques: a Time Series Analysis (TSA), based on the evaluation of trend and periodicity of the series, and a Regression (REGR) method, based on a modification of linear stochastic regression, will be presented and compared. The error analysis will show interesting features of both the models. In addition, the differences between a deterministic (TSA) and a stochastic imputation approach will be highlighted in terms of dataset mean and variance preservation.