The dangerous effect of acoustical noise on health, both auditory and non-auditory, is largely documented in literature. The study of noise problem and the development of several predictive models are very important issues in cities and areas in which relevant noise sources are present. A model based on threshold exceedances is suitable since a great number of national regulations define limits to acoustical levels, according to the area and to the kind of buildings and activities occurring in them. In this paper, a non-homogeneous Poisson model will be presented and applied to a large dataset of noise measurements. The aim of this model is to predict, from a probabilistic point of view, the number of threshold exceedances in a given future period. This can be achieved estimating the parameters of the mean function, that represents the number of surpassings of the threshold as a function of the time, and of the rate function of the Poisson process. Different models of the rate function form can be chosen, according to their functional dependence from time. The choice adopted in this work will be the Goel Okumoto (GO) model. The estimation of the parameters probability distributions will be performed using a Bayesian approach, based on Monte Carlo Markov Chains and Gibbs algorithm. Once the parameters will be tuned, the mean function can be compared with the observed exceedances plot. The two GO parameters will be estimated using non-informative uniform prior distributions and also informative gamma distributions. Different starting points sets will be implemented for the Markov chains to evaluate possible effects on the posterior distributions.

Environmental noise level threshold surpassing analysis by non-homogeneous Poisson model with informative and non-informative prior distributions

GUARNACCIA, CLAUDIO;QUARTIERI, Joseph;TEPEDINO, CARMINE
2016-01-01

Abstract

The dangerous effect of acoustical noise on health, both auditory and non-auditory, is largely documented in literature. The study of noise problem and the development of several predictive models are very important issues in cities and areas in which relevant noise sources are present. A model based on threshold exceedances is suitable since a great number of national regulations define limits to acoustical levels, according to the area and to the kind of buildings and activities occurring in them. In this paper, a non-homogeneous Poisson model will be presented and applied to a large dataset of noise measurements. The aim of this model is to predict, from a probabilistic point of view, the number of threshold exceedances in a given future period. This can be achieved estimating the parameters of the mean function, that represents the number of surpassings of the threshold as a function of the time, and of the rate function of the Poisson process. Different models of the rate function form can be chosen, according to their functional dependence from time. The choice adopted in this work will be the Goel Okumoto (GO) model. The estimation of the parameters probability distributions will be performed using a Bayesian approach, based on Monte Carlo Markov Chains and Gibbs algorithm. Once the parameters will be tuned, the mean function can be compared with the observed exceedances plot. The two GO parameters will be estimated using non-informative uniform prior distributions and also informative gamma distributions. Different starting points sets will be implemented for the Markov chains to evaluate possible effects on the posterior distributions.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11386/4667653
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? ND
social impact