On the fuzzy Bayesian inference of population annoyance level caused by noise exposure

• Authors: Juan-Miguel León-Rojas;Valentín Masero;Montaña Morales
• Venue: Proceedings of the 2003 ACM symposium on Applied computing
• Year: 2003

Too many people suffer from noise levels that scientists and health experts consider to be unacceptable, where most people become annoyed, where sleep is disturbed and where adverse health effects are to be feared. The present paper centers on inferring individuals annoyance level caused by noise exposure. The starting point of our thoughts is the impossibility of assigning an exact number to observed values. Observed sound level values are imprecise. Although sound level meters are more and more accurate, they cannot assign yet an exact number to an observed value. Really, we assume that actually, no exact number can be assigned to a sound level observed value. If our interest lies in maximum effects of the propagation of this imprecision, an interval-based approach seems to be adequate. Anyway, although in a nearly future we could fill these gaps, it seems better to think about a time interval and its associated data-interval corresponding to the observed sound level range. Although we revised a maximum likelihood solution, the aim of this paper is to discuss how imprecision of observation values is propagated when fuzzy Bayesian inference for these non-precise observations is carried out.