Markov chain modeling for air pollution index based on maximum a posteriori method

Yousif Alyousifi, Kamarulzaman Ibrahim, Wei Kang, Wan Zawiah Wan Zin

Research output: Contribution to journalArticle

Abstract

Air pollution is a major environmental problem, which brings about a threat to human health and the natural environment. Thus, determination and assessment of the level of air pollution is an important component in monitoring of the air quality. This study involves estimating the transition probability matrix of the Markov chain model based on maximum a posteriori (MAP) method using hourly data of air pollution index (API). The API data has been collected from seven air-monitoring stations in peninsular Malaysia. The estimated transition probability matrix is used to determine the characteristics of air pollution such as the steady-state probability and the expected return time. The transition probabilities of the Markov chain are fitted based on maximum a posteriori method under three different priors, which are Dirichlet, Jeffreys, and uniform. The results found show that the maximum a posteriori method under the Dirichlet prior produced the most precise estimates as compared with the other priors. In addition, for the areas considered in the study, the moderate state is more persistent as opposed to unhealthy states indicating that the problem of air quality is not very serious. In general, this study could provide important implications for developing proper strategies for managing air quality and for improving public health.

Original languageEnglish
Pages (from-to)1521-1531
Number of pages11
JournalAir Quality, Atmosphere and Health
Volume12
Issue number12
DOIs
Publication statusPublished - 1 Dec 2019

Fingerprint

Markov Chains
Air Pollution
Markov chain
Air pollution
Markov processes
atmospheric pollution
Air
Air quality
air quality
modeling
matrix
Monitoring
Malaysia
Public health
public health
Public Health
index
method
Health
air

Keywords

  • Air pollution
  • Bayesian inference
  • Dirichlet distribution
  • Markov chain
  • Maximum a posteriori estimation
  • Steady-state probability

ASJC Scopus subject areas

  • Pollution
  • Atmospheric Science
  • Management, Monitoring, Policy and Law
  • Health, Toxicology and Mutagenesis

Cite this

Markov chain modeling for air pollution index based on maximum a posteriori method. / Alyousifi, Yousif; Ibrahim, Kamarulzaman; Kang, Wei; Zin, Wan Zawiah Wan.

In: Air Quality, Atmosphere and Health, Vol. 12, No. 12, 01.12.2019, p. 1521-1531.

Research output: Contribution to journalArticle

@article{44463e336f954b89b4cd972c434bf336,
title = "Markov chain modeling for air pollution index based on maximum a posteriori method",
abstract = "Air pollution is a major environmental problem, which brings about a threat to human health and the natural environment. Thus, determination and assessment of the level of air pollution is an important component in monitoring of the air quality. This study involves estimating the transition probability matrix of the Markov chain model based on maximum a posteriori (MAP) method using hourly data of air pollution index (API). The API data has been collected from seven air-monitoring stations in peninsular Malaysia. The estimated transition probability matrix is used to determine the characteristics of air pollution such as the steady-state probability and the expected return time. The transition probabilities of the Markov chain are fitted based on maximum a posteriori method under three different priors, which are Dirichlet, Jeffreys, and uniform. The results found show that the maximum a posteriori method under the Dirichlet prior produced the most precise estimates as compared with the other priors. In addition, for the areas considered in the study, the moderate state is more persistent as opposed to unhealthy states indicating that the problem of air quality is not very serious. In general, this study could provide important implications for developing proper strategies for managing air quality and for improving public health.",
keywords = "Air pollution, Bayesian inference, Dirichlet distribution, Markov chain, Maximum a posteriori estimation, Steady-state probability",
author = "Yousif Alyousifi and Kamarulzaman Ibrahim and Wei Kang and Zin, {Wan Zawiah Wan}",
year = "2019",
month = "12",
day = "1",
doi = "10.1007/s11869-019-00764-y",
language = "English",
volume = "12",
pages = "1521--1531",
journal = "Air Quality, Atmosphere and Health",
issn = "1873-9318",
publisher = "Springer Netherlands",
number = "12",

}

TY - JOUR

T1 - Markov chain modeling for air pollution index based on maximum a posteriori method

AU - Alyousifi, Yousif

AU - Ibrahim, Kamarulzaman

AU - Kang, Wei

AU - Zin, Wan Zawiah Wan

PY - 2019/12/1

Y1 - 2019/12/1

N2 - Air pollution is a major environmental problem, which brings about a threat to human health and the natural environment. Thus, determination and assessment of the level of air pollution is an important component in monitoring of the air quality. This study involves estimating the transition probability matrix of the Markov chain model based on maximum a posteriori (MAP) method using hourly data of air pollution index (API). The API data has been collected from seven air-monitoring stations in peninsular Malaysia. The estimated transition probability matrix is used to determine the characteristics of air pollution such as the steady-state probability and the expected return time. The transition probabilities of the Markov chain are fitted based on maximum a posteriori method under three different priors, which are Dirichlet, Jeffreys, and uniform. The results found show that the maximum a posteriori method under the Dirichlet prior produced the most precise estimates as compared with the other priors. In addition, for the areas considered in the study, the moderate state is more persistent as opposed to unhealthy states indicating that the problem of air quality is not very serious. In general, this study could provide important implications for developing proper strategies for managing air quality and for improving public health.

AB - Air pollution is a major environmental problem, which brings about a threat to human health and the natural environment. Thus, determination and assessment of the level of air pollution is an important component in monitoring of the air quality. This study involves estimating the transition probability matrix of the Markov chain model based on maximum a posteriori (MAP) method using hourly data of air pollution index (API). The API data has been collected from seven air-monitoring stations in peninsular Malaysia. The estimated transition probability matrix is used to determine the characteristics of air pollution such as the steady-state probability and the expected return time. The transition probabilities of the Markov chain are fitted based on maximum a posteriori method under three different priors, which are Dirichlet, Jeffreys, and uniform. The results found show that the maximum a posteriori method under the Dirichlet prior produced the most precise estimates as compared with the other priors. In addition, for the areas considered in the study, the moderate state is more persistent as opposed to unhealthy states indicating that the problem of air quality is not very serious. In general, this study could provide important implications for developing proper strategies for managing air quality and for improving public health.

KW - Air pollution

KW - Bayesian inference

KW - Dirichlet distribution

KW - Markov chain

KW - Maximum a posteriori estimation

KW - Steady-state probability

UR - http://www.scopus.com/inward/record.url?scp=85076099427&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85076099427&partnerID=8YFLogxK

U2 - 10.1007/s11869-019-00764-y

DO - 10.1007/s11869-019-00764-y

M3 - Article

AN - SCOPUS:85076099427

VL - 12

SP - 1521

EP - 1531

JO - Air Quality, Atmosphere and Health

JF - Air Quality, Atmosphere and Health

SN - 1873-9318

IS - 12

ER -