Comparing neural networks and regression models for ozone forecasting

Research output: Contribution to journalArticle

327 Citations (Scopus)

Abstract

Many large metropolitan areas experience elevated concentrations of ground-level ozone pollution during the summertime 'smog season'. Local environmental or health agencies often need to make daily air pollution forecasts for public advisories and for input into decisions regarding abatement measures and air quality management. Such forecasts are usually based on statistical relationships between weather conditions and ambient air pollution concentrations. Multivariate linear regression models have been widely used for this purpose, and well-specified regressions can provide reasonable results. However, pollution-weather relationships are typically complex and nonlinear-especially for ozone-properties that might be better captured by neural networks. This study investigates the potential for using neural networks to forecast ozone pollution, as compared to traditional regression models. Multiple regression models and neural networks are examined for a range of cities under different climate and ozone regimes, enabling a comparative study of the two approaches. Model comparison statistics indicate that neutral network techniques are somewhat (but not dramatically) better than regression models for daily ozone prediction, and that all types of models are sensitive to different weather-zone regimes and the role of persistence in aiding predictions.

Original languageEnglish (US)
Pages (from-to)653-663
Number of pages11
JournalJournal of the Air and Waste Management Association
Volume47
Issue number6
StatePublished - 1997

Fingerprint

Ozone
ozone
Neural networks
Pollution
Air pollution
pollution
atmospheric pollution
weather
smog
Quality management
prediction
Air quality
Linear regression
ambient air
metropolitan area
multiple regression
comparative study
persistence
Health
Statistics

ASJC Scopus subject areas

  • Atmospheric Science
  • Waste Management and Disposal
  • Environmental Engineering
  • Environmental Science(all)
  • Environmental Chemistry

Cite this

Comparing neural networks and regression models for ozone forecasting. / Comrie, Andrew.

In: Journal of the Air and Waste Management Association, Vol. 47, No. 6, 1997, p. 653-663.

Research output: Contribution to journalArticle

@article{b38f875b0432422ca6e218b2c5492b4c,
title = "Comparing neural networks and regression models for ozone forecasting",
abstract = "Many large metropolitan areas experience elevated concentrations of ground-level ozone pollution during the summertime 'smog season'. Local environmental or health agencies often need to make daily air pollution forecasts for public advisories and for input into decisions regarding abatement measures and air quality management. Such forecasts are usually based on statistical relationships between weather conditions and ambient air pollution concentrations. Multivariate linear regression models have been widely used for this purpose, and well-specified regressions can provide reasonable results. However, pollution-weather relationships are typically complex and nonlinear-especially for ozone-properties that might be better captured by neural networks. This study investigates the potential for using neural networks to forecast ozone pollution, as compared to traditional regression models. Multiple regression models and neural networks are examined for a range of cities under different climate and ozone regimes, enabling a comparative study of the two approaches. Model comparison statistics indicate that neutral network techniques are somewhat (but not dramatically) better than regression models for daily ozone prediction, and that all types of models are sensitive to different weather-zone regimes and the role of persistence in aiding predictions.",
author = "Andrew Comrie",
year = "1997",
language = "English (US)",
volume = "47",
pages = "653--663",
journal = "Journal of the Air and Waste Management Association",
issn = "1096-2247",
publisher = "Air and Waste Management Association",
number = "6",

}

TY - JOUR

T1 - Comparing neural networks and regression models for ozone forecasting

AU - Comrie, Andrew

PY - 1997

Y1 - 1997

N2 - Many large metropolitan areas experience elevated concentrations of ground-level ozone pollution during the summertime 'smog season'. Local environmental or health agencies often need to make daily air pollution forecasts for public advisories and for input into decisions regarding abatement measures and air quality management. Such forecasts are usually based on statistical relationships between weather conditions and ambient air pollution concentrations. Multivariate linear regression models have been widely used for this purpose, and well-specified regressions can provide reasonable results. However, pollution-weather relationships are typically complex and nonlinear-especially for ozone-properties that might be better captured by neural networks. This study investigates the potential for using neural networks to forecast ozone pollution, as compared to traditional regression models. Multiple regression models and neural networks are examined for a range of cities under different climate and ozone regimes, enabling a comparative study of the two approaches. Model comparison statistics indicate that neutral network techniques are somewhat (but not dramatically) better than regression models for daily ozone prediction, and that all types of models are sensitive to different weather-zone regimes and the role of persistence in aiding predictions.

AB - Many large metropolitan areas experience elevated concentrations of ground-level ozone pollution during the summertime 'smog season'. Local environmental or health agencies often need to make daily air pollution forecasts for public advisories and for input into decisions regarding abatement measures and air quality management. Such forecasts are usually based on statistical relationships between weather conditions and ambient air pollution concentrations. Multivariate linear regression models have been widely used for this purpose, and well-specified regressions can provide reasonable results. However, pollution-weather relationships are typically complex and nonlinear-especially for ozone-properties that might be better captured by neural networks. This study investigates the potential for using neural networks to forecast ozone pollution, as compared to traditional regression models. Multiple regression models and neural networks are examined for a range of cities under different climate and ozone regimes, enabling a comparative study of the two approaches. Model comparison statistics indicate that neutral network techniques are somewhat (but not dramatically) better than regression models for daily ozone prediction, and that all types of models are sensitive to different weather-zone regimes and the role of persistence in aiding predictions.

UR - http://www.scopus.com/inward/record.url?scp=0031172117&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031172117&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0031172117

VL - 47

SP - 653

EP - 663

JO - Journal of the Air and Waste Management Association

JF - Journal of the Air and Waste Management Association

SN - 1096-2247

IS - 6

ER -