Pelliccioni, Armando. and Tirabassi, T. (2008) Air pollution model and neural network: An integrated modelling system. Il nuovo cimento C, 31 (3). pp. 253-273. ISSN 1826-9885
|
Text
ncc9295.pdf - Accepted Version Download (602kB) | Preview |
Abstract
It is well known that neural networks can work as universal approximators of non-linear functions and they have become a useful tool either where any precise phenomenological model is available or when uncertainty complicates the application of deterministic modelling as, for example, in environmental systems. Usually, NN models are using as regression tool. We have developed an integrated modelling system coupling an air dispersion model with a neural network method both to simulate the influence of important parameters on air pollution models and to minimize the input neural net variables. In our approach, an optimised 3-Layer Perception is used to filter the air pollution concentrations evaluated by means of the non-Gaussian analytical model ADMD. We applied this methodology to the wellknown Indianapolis urban data set which deals with a release of pollutants from an elevated emission source.
Item Type: | Article |
---|---|
Uncontrolled Keywords: | Air quality and air pollution ; Instruments for environmental pollution measurements ; Neural networks |
Subjects: | 300 Scienze sociali > 360 Problemi e servizi sociali; associazioni > 363 Altri problemi e servizi sociali > 363.7 Problemi ambientali (classificare qui la tutela ambientale; l’effetto dei rifiuti, dell’inquinamento, delle iniziative per controllarli) > 363.73 Inquinamento > 363.739 Inquinamento di specifici ambienti > 363.7392 Inquinamento atmosferico |
Depositing User: | Marina Spanti |
Date Deposited: | 23 Mar 2020 15:37 |
Last Modified: | 23 Mar 2020 15:37 |
URI: | http://eprints.bice.rm.cnr.it/id/eprint/16357 |
Actions (login required)
View Item |