NN-AirPol

A neural-networks-based method for air pollution evaluation and control

Ferhat Karaca, Alexander Nikov, Omar Alagha

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

A method for air pollution evaluation and control, based on one of the most popular neural networks - the backpropagation algorithm, is proposed. After the backpropagation training, the neural network, based on weather forecasting data, determines the future concentration of critical air pollution indicators. Depending on these concentrations, relevant episode warnings and actions are activated. A case study is carried out to illustrate and validate the method proposed, based on Istanbul air pollution data. Sulphur dioxide and inhalable participate matter are selected as air pollution indicators (neural network outputs). Relevant episode measures are proposed. Among ten backpropagation algorithms, the BFGS algorithm (Quasi-Newton algorithms) is adopted since it showed the lowest training error. A comparison of NN-AirPol method against regression and perceptron models showed significantly better performance.

Original languageEnglish
Pages (from-to)310-325
Number of pages16
JournalInternational Journal of Environment and Pollution
Volume28
Issue number3-4
DOIs
Publication statusPublished - Dec 18 2006
Externally publishedYes

Fingerprint

Air pollution
atmospheric pollution
Neural networks
Backpropagation algorithms
Weather forecasting
Sulfur Dioxide
weather forecasting
Sulfur dioxide
Backpropagation
sulfur dioxide
evaluation
method
pollution indicator

Keywords

  • Air pollution
  • Backpropagation algorithm
  • Modelling
  • Optimisation

ASJC Scopus subject areas

  • Environmental Science(all)
  • Environmental Chemistry
  • Pollution

Cite this

NN-AirPol : A neural-networks-based method for air pollution evaluation and control. / Karaca, Ferhat; Nikov, Alexander; Alagha, Omar.

In: International Journal of Environment and Pollution, Vol. 28, No. 3-4, 18.12.2006, p. 310-325.

Research output: Contribution to journalArticle

@article{17687373cc7344ee822f24337af20e41,
title = "NN-AirPol: A neural-networks-based method for air pollution evaluation and control",
abstract = "A method for air pollution evaluation and control, based on one of the most popular neural networks - the backpropagation algorithm, is proposed. After the backpropagation training, the neural network, based on weather forecasting data, determines the future concentration of critical air pollution indicators. Depending on these concentrations, relevant episode warnings and actions are activated. A case study is carried out to illustrate and validate the method proposed, based on Istanbul air pollution data. Sulphur dioxide and inhalable participate matter are selected as air pollution indicators (neural network outputs). Relevant episode measures are proposed. Among ten backpropagation algorithms, the BFGS algorithm (Quasi-Newton algorithms) is adopted since it showed the lowest training error. A comparison of NN-AirPol method against regression and perceptron models showed significantly better performance.",
keywords = "Air pollution, Backpropagation algorithm, Modelling, Optimisation",
author = "Ferhat Karaca and Alexander Nikov and Omar Alagha",
year = "2006",
month = "12",
day = "18",
doi = "10.1504/IJEP.2006.011214",
language = "English",
volume = "28",
pages = "310--325",
journal = "International Journal of Environment and Pollution",
issn = "0957-4352",
publisher = "Inderscience Enterprises Ltd",
number = "3-4",

}

TY - JOUR

T1 - NN-AirPol

T2 - A neural-networks-based method for air pollution evaluation and control

AU - Karaca, Ferhat

AU - Nikov, Alexander

AU - Alagha, Omar

PY - 2006/12/18

Y1 - 2006/12/18

N2 - A method for air pollution evaluation and control, based on one of the most popular neural networks - the backpropagation algorithm, is proposed. After the backpropagation training, the neural network, based on weather forecasting data, determines the future concentration of critical air pollution indicators. Depending on these concentrations, relevant episode warnings and actions are activated. A case study is carried out to illustrate and validate the method proposed, based on Istanbul air pollution data. Sulphur dioxide and inhalable participate matter are selected as air pollution indicators (neural network outputs). Relevant episode measures are proposed. Among ten backpropagation algorithms, the BFGS algorithm (Quasi-Newton algorithms) is adopted since it showed the lowest training error. A comparison of NN-AirPol method against regression and perceptron models showed significantly better performance.

AB - A method for air pollution evaluation and control, based on one of the most popular neural networks - the backpropagation algorithm, is proposed. After the backpropagation training, the neural network, based on weather forecasting data, determines the future concentration of critical air pollution indicators. Depending on these concentrations, relevant episode warnings and actions are activated. A case study is carried out to illustrate and validate the method proposed, based on Istanbul air pollution data. Sulphur dioxide and inhalable participate matter are selected as air pollution indicators (neural network outputs). Relevant episode measures are proposed. Among ten backpropagation algorithms, the BFGS algorithm (Quasi-Newton algorithms) is adopted since it showed the lowest training error. A comparison of NN-AirPol method against regression and perceptron models showed significantly better performance.

KW - Air pollution

KW - Backpropagation algorithm

KW - Modelling

KW - Optimisation

UR - http://www.scopus.com/inward/record.url?scp=33845414100&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33845414100&partnerID=8YFLogxK

U2 - 10.1504/IJEP.2006.011214

DO - 10.1504/IJEP.2006.011214

M3 - Article

VL - 28

SP - 310

EP - 325

JO - International Journal of Environment and Pollution

JF - International Journal of Environment and Pollution

SN - 0957-4352

IS - 3-4

ER -