Prediction of airport acoustical noise by deterministic decomposition and seasonal arima techniques

Abstract

In this paper a time series of hourly equivalent noise levels acquired near the international airport of Nice, France, is analysed. Two different techniques are proposed to model and forecast the time series: deterministic decomposition and seasonal autoregressive moving average. The two models are defined and fitted on the calibration dataset. Subsequently, the developed models are tested comparing their forecasts with 25 noise level data not used in the calibration phase. A detailed error analysis, by means of statistics and metrics, will be presented to test the models performances.