Bayesian Information Criterion for Fitting the Optimum Order of Markov Chain Models: Methodology and Application to Air Pollution Data
Abstract
:1. Introduction
2. Methodology
2.1. Discrete-Time Markov Chain Model
2.2. Testing the Markov Property and Divergence for Empirically Estimated Transition Matrix of MC Sequence
2.3. Fitting the Optimum Order of the MC Model
3. Results and Discussion
3.1. Application to Air Pollution Data
3.2. Trend Analysis for the API Data
3.3. The Performance of the Markov Chain Modeling
3.4. Assessment of the Markov Property and Divergence of Markov Chain Sequence
3.5. Investigating the Optimum Order of Markov Chain Models
4. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Manisalidis, I.; Elisavet, S.; Agathangelos, S.; Eugenia, B. Environmental and health impacts of air pollution: A review. Front. Public Health 2020, 8, 14. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Puentes, R.; Marchant, C.; Leiva, V.; Figueroa-Zúñiga, J.I.; Ruggeri, F. Predicting PM2.5 and PM10 levels during critical episodes management in Santiago, Chile, with a bivariate Birnbaum-Saunders log-linear model. Mathematics 2021, 9, 645. [Google Scholar] [CrossRef]
- Othman, J.; Sahani, M.; Mahmud, M.; Ahmad, M.K.S. Transboundary smoke haze pollution in Malaysia: Inpatient health impacts and economic valuation. Environ. Pollut. 2014, 189, 194–201. [Google Scholar] [CrossRef]
- Gass, K.; Klein, M.; Sarnat, S.E.; Winquist, A.; Darrow, L.A.; Flanders, W.D.; Chang, H.H.; Mulholland, J.A.; Tolbert, P.E.; Strickland, M.J. Associations between ambient air pollutant mixtures and pediatric asthma emergency department visits in three cities: A classification and regression tree approach. Environ. Health 2015, 14, 1–14. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Sun, D.; Li, X. Application of Markov chain model on environmental fate of phenanthrene in soil and groundwater. Procedia Environ. Sci. 2010, 2, 814–823. [Google Scholar] [CrossRef] [Green Version]
- Zhou, Y.; Wang, L.; Zhong, R.; Tan, Y. A Markov chain based demand prediction model for stations in bike sharing systems. Math. Probl. Eng. 2018, 2018, 1–8. [Google Scholar] [CrossRef] [Green Version]
- Choji, D.N.; Eduno, S.N.; Kassem, G.T. Markov chain model application on share price movement in stock market. Comput. Eng. Intell. Syst. 2013, 4, 84–95. [Google Scholar]
- Saad, S.A.; Adnan, F.A.; Ibrahim, H.; Rahim, R. Manpower planning using Markov Chain model. In Proceedings of the AIP Conference Proceedings, Penang, Malaysia, 6–8 November 2013; AIP: College Park, MD, USA, 2014; pp. 1123–1127. [Google Scholar]
- Salas-Páez, C.; Quintana-Romero, L.; Mendoza-González, M.A.; Álvarez-García, J. Analysis of Job Transitions in Mexico with Markov Chains in Discrete Time. Mathematics 2022, 10, 1693. [Google Scholar] [CrossRef]
- Sahin, A.D.; Sen, Z. First-order Markov chain approach to wind speed modelling. J. Wind. Eng. Ind. Aerodyn. 2001, 89, 263–269. [Google Scholar] [CrossRef]
- Shamshad, A.; Bawadi, M.; Hussin, W.W.; Majid, T.A.; Sanusi, S. First and second order Markov chain models for synthetic generation of wind speed time series. Energy 2005, 30, 693–708. [Google Scholar] [CrossRef]
- Fawcett, L.; Walshaw, D. Markov chain models for extreme wind speeds. Env. Off. J. Int. Env. Soc. 2006, 17, 795–809. [Google Scholar] [CrossRef]
- Carpinone, A.; Giorgio, M.; Langella, R.; Testa, A. Markov chain modeling for very-short-term wind power forecasting. Electr. Power Syst. Res. 2015, 122, 152–158. [Google Scholar] [CrossRef] [Green Version]
- Larsen, L.; Bradley, R.; Honcoop, G. A new method of characterizing the variability of air quality-related indicators. In Proceedings of the Air and waste management association’s international specialty conference of tropospheric ozone and the environment, Los Angeles, CA, USA, 3 May 1990. [Google Scholar]
- Hoyos, L.; Lara, P.; Ortiz, E.; Bracho, R.L.; González, J. Evaluation of air pollution control policies in Mexico City using finite Markov chain observation model. Rev. Matemática Teoría Apl. 2009, 16, 255–266. [Google Scholar] [CrossRef] [Green Version]
- Rodrigues, E.R.; Achcar, J.A. Applications of Discrete-Time Markov Chains and Poisson Processes to Air Pollution Modeling and Studies; Springer Science & Business Media: Berlin/Heidelberg, Germany, 2012. [Google Scholar]
- Asadollahfardi, G.; Zangooei, H.; Aria, S.H. Predicting PM 2.5 concentrations using artificial neural networks and Markov chain, a case study Karaj City. Asian J. Atmos. Environ. 2016, 10, 67–79. [Google Scholar] [CrossRef] [Green Version]
- Nebenzal, A.; Fishbain, B. Long-term forecasting of nitrogen dioxide ambient levels in metropolitan areas using the discrete-time Markov model. Environ. Model. Softw. 2018, 107, 175–185. [Google Scholar] [CrossRef]
- Mohamad, N.; Deni, S.; Ul-Saufie, A. Application of the First Order of Markov Chain Model in Describing the PM10 Occurrences in Shah Alam and Jerantut, Malaysia. Pertanika J. Sci. Technol. 2018, 26, 367–378. [Google Scholar]
- Alyousifi, Y.; Masseran, N.; Ibrahim, K. Modeling the stochastic dependence of air pollution index data. Stoch. Environ. Res. Risk Assess. 2018, 32, 1603–1611. [Google Scholar] [CrossRef]
- Alyousifi, Y.; Ibrahim, K.; Kang, W.; Zin, W.Z.W. Markov chain modeling for air pollution index based on maximum a posteriori method. Air Qual. Atmos. Health 2019, 12, 1521–1531. [Google Scholar] [CrossRef]
- Chen, J.-C.; Wu, Y.J. Discrete-time Markov chain for prediction of air quality index. J. Ambient. Intell. Humaniz. Comput. 2020, 9, 1–10. [Google Scholar] [CrossRef]
- Alyousifi, Y.; Ibrahim, K.; Kang, W.; Zin, W.Z.W. Modeling the spatio-temporal dynamics of air pollution index based on spatial Markov chain model. Environ. Monit. Assess. 2020, 192, 1–24. [Google Scholar]
- Tijms, H.C. A First Course in Stochastic Models; John Wiley and Sons: Hoboken, NJ, USA, 2003. [Google Scholar]
- Ching, W.-K.; Huang, X.; Ng, M.K.; Siu, T.-K. Higher-order markov chains. In Markov Chains; Springer: Berlin/Heidelberg, Germany, 2013; pp. 141–176. [Google Scholar]
- Ibe, O. Markov Processes for Stochastic Modeling; Newnes: Oxford, UK, 2013. [Google Scholar]
- Privault, N. Understanding markov chains. Ex. Appl. Publ. 2013, 357, 358. [Google Scholar]
- Grinstead, C.M.; Snell, J.L. Introduction to Probability; American Mathematical Soc.: Providence, RI, USA, 1997. [Google Scholar]
- Nikolić, D.; Milošević, N.; Mihajlović, I.; Živković, Ž.; Tasić, V.; Kovačević, R.; Petrović, N. Multi-criteria analysis of air pollution with SO2 and PM10 in urban area around the copper smelter in Bor, Serbia. Water Air Soil Pollut. 2010, 206, 369–383. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Li, C.-K.; Zhang, S. Stationary probability vectors of higher-order Markov chains. Linear Algebra Its Appl. 2015, 473, 114–125. [Google Scholar] [CrossRef]
- Anderson, T.W.; Goodman, L.A. Statistical inference about Markov chains. Ann. Math. Stat. 1957, 28, 89–110. [Google Scholar] [CrossRef]
- Billingsley, P. Statistical methods in Markov chains. Ann. Math. Stat. 1961, 32, 12–40. [Google Scholar] [CrossRef]
- Katz, R.W. On some criteria for estimating the order of a Markov chain. Technometrics 1981, 23, 243–249. [Google Scholar] [CrossRef]
- Deni, S.M.; Jemain, A.A.; Ibrahim, K. Fitting optimum order of Markov chain models for daily rainfall occurrences in Peninsular Malaysia. Theor. Appl. Climatol. 2009, 97, 109–121. [Google Scholar] [CrossRef]
- Bickenbach, F.; Bode, E. Evaluating the Markov property in studies of economic convergence. Int. Reg. Sci. Rev. 2003, 26, 363–392. [Google Scholar] [CrossRef]
- Skuriat-Olechnowska, M. Statistical inference and hypothesis testing for Markov chains with Interval Censoring. Master’s Thesis, Delft University of Technology, Delft, The Netherlands, 2005. [Google Scholar]
- Spedicato, G. A package to handle and analyse discrete time Markov Chains. R Package Version 2014, 95, 1–17. [Google Scholar]
- Bickenbach, F.; Bode, E. Markov or Not Markov-This Should Be a Question; Kiel working paper; Kiel Institute of World Economics: Kiel, Germany, 2001. [Google Scholar]
- Schoof, J.; Pryor, S. On the proper order of Markov chain model for daily precipitation occurrence in the contiguous United States. J. Appl. Meteorol. Climatol. 2008, 47, 2477–2486. [Google Scholar] [CrossRef]
- Jimoh, O.; Webster, P. The optimum order of a Markov chain model for daily rainfall in Nigeria. J. Hydrol. 1996, 185, 45–69. [Google Scholar] [CrossRef]
- Mohamad, N.; Deni, S.; Japeri, A. Modeling of daily PM10 concentration occurrence using Markov Chain model in Shah Alam, Malaysia. J. Environ. Sci. Technol. 2017, 10, 96–106. [Google Scholar] [CrossRef] [Green Version]
- Dastidar, A.G.; Ghosh, D.; Dasgupta, S.; De, U. Higher order Markov chain models for monsoon rainfall over West Bengal, India. Indian J. Radio Space Phys. 2010, 39, 39–44. [Google Scholar]
- Wilks, D.S. Statistical Methods in the Atmospheric Sciences; Academic Press: London, UK, 2011; Volume 100. [Google Scholar]
- Bowerman, B.L.; O’Connell, R.T.; Koehler, A.B. Forecasting, Time Series, and Regression: An Applied Approach; South-Western Pub: San Diego, CA, USA, 2005; Volume 4. [Google Scholar]
- Masseran, N.; Safari, M.A.M. Modeling the transition behaviors of PM10 pollution index. Environ. Monit. Assess. 2020, 192, 1–15. [Google Scholar] [CrossRef] [PubMed]
- Chin, S.A. A relativistic many-body theory of high density matter. Ann. Phys. 1977, 108, 301–367. [Google Scholar] [CrossRef]
- World Health Organization. Air Quality Guidelines for Europe; World Health Organization, Regional Office for Europe: Geneva, Switzerland, 2000. [Google Scholar]
- Barbu, V.S.; Bérard, C.; Cellier, D.; Sautreuil, M.; Vergne, N. SMM: An R Package for Estimation and Simulation of Discrete-time semi-Markov Models. R J. 2018, 10, 226. [Google Scholar] [CrossRef] [Green Version]
- Vergne, N. Drifting Markov models with polynomial drift and applications to DNA sequences. Stat. Appl. Genet. Mol. Biol. 2008, 7, 6. [Google Scholar] [CrossRef] [Green Version]
- Barbu, V.S.; Vergne, N. Reliability and survival analysis for drifting Markov models: Modeling and estimation. Methodol. Comput. Appl. Probab. 2019, 21, 1407–1429. [Google Scholar] [CrossRef] [Green Version]
Station | Area | Statistics | |||||||
---|---|---|---|---|---|---|---|---|---|
Max | Mean | Standard Deviation | Prop. API > 100 | Prop. API > 200 | Prop. API > 300 | Kurtosis | Skewness | ||
A1 | Kuala Lumpur | 203 | 46.91 | 19.05 | 0.0146 | 0.00009 | 0.0000 | 7.8650 | 1.84787 |
A2 | Klang | 495 | 56.90 | 25.13 | 0.0333 | 0.00372 | 0.0024 | 90.333 | 7.18003 |
A3 | Cheras | 201 | 49.85 | 17.25 | 0.0141 | 0.00003 | 0.0000 | 6.3822 | 1.58419 |
A4 | Petaling Jaya | 231 | 47.37 | 16.53 | 0.0106 | 0.00061 | 0.0000 | 19.469 | 2.88533 |
A5 | Banting | 323 | 51.35 | 19.73 | 0.0220 | 0.00182 | 0.0005 | 35.720 | 1.58419 |
A6 | Shah Alam | 301 | 47.32 | 18.41 | 0.0146 | 0.0008 | 0.00004 | 20.606 | 2.72789 |
A7 | Kuala Selangor | 247 | 40.16 | 19.09 | 0.0135 | 0.00057 | 0.0000 | 13.121 | 2.47169 |
Station | Verify Markov Property | Verify Divergence | ||||
---|---|---|---|---|---|---|
Statistics | df | p-Value | Statistics | df | p-Value | |
Kuala Lumpur | 201.939 | 27 | 0.00002 | 1.908207 | 6 | 0.75264 |
Klang | 47.5692 | 27 | 0.0086 | 0.010878 | 6 | 0.99857 |
Cheras | 12.6243 | 27 | 0.9914 | 0.001336 | 6 | 0.9999 |
Petaling Jaya | 18.8424 | 27 | 0.8856 | 0.009457 | 6 | 0.99998 |
Banting | 54.1234 | 27 | 0.0015 | 0.066857 | 6 | 0.15348 |
Shah Alam | 77.0592 | 27 | 0.00001 | 8.590637 | 6 | 0.07219 |
Kuala Selangor | 0.04123 | 27 | 1.0000 | 0.036617 | 6 | 0.99983 |
Station Order | BIC of the Hourly API | |||||||
---|---|---|---|---|---|---|---|---|
KL | Klang | Cheras | PJ | Banting | ShA | KS | ||
Observed API | 1 | 1696.011 | 788.8619 | 1729.459 | 472.9546 | 1161.643 | 1084.574 | 374.2396 |
2 | 1468.189 | 619.6946 | 1539.279 | 485.638 | 990.832 | 994.5611 | 381.2125 | |
3 | 1350.565 | 671.6492 | 1238.221 | 690.3285 | 967.8343 | 1052.922 | 578.222 | |
4 | 2380.511 | 1772.959 | 2338.917 | 1791.672 | 2069.016 | 2141.571 | 1679.577 | |
5 | 5668.314 | 5072.603 | 5637.949 | 5091.351 | 5363.804 | 5441.096 | 4979.267 | |
6 | 15,562.45 | 14,967.25 | 15,531.98 | 14986.03 | 15,245.63 | 15,335.63 | 14,873.96 | |
7 | 45,241.6 | 44,646.92 | 45,211.03 | 44,665.73 | 44,925.17 | 45,015.18 | 44,553.67 | |
8 | 134,275.8 | 133,681.5 | 134,245.1 | 133,700.5 | 133,959.7 | 134,049.8 | 133,588.4 | |
9 | 401,375.1 | 400,781.4 | 401,344.3 | 400,800.3 | 401,059.4 | 401,149.5 | 400,688.3 | |
10 | 1,202,670 | 1,202,077 | 1,202,639 | 1,202,096 | 1,202,355 | 1,202,445 | 1,201,984 | |
Simulated API | 1 | 1842.167 | 712.794 | 1803.19 | 550.6943 | 1076.533 | 1080.534 | 500.2132 |
2 | 1614.669 | 610.7687 | 1588.566 | 517.9346 | 891.2364 | 922.0258 | 443.4873 | |
3 | 1103.086 | 651.6381 | 1115.187 | 588.1897 | 691.0557 | 689.9981 | 578.0087 | |
4 | 2203.769 | 1752.948 | 2215.855 | 1689.53 | 1792.237 | 1791.181 | 1679.361 | |
5 | 5502.787 | 5052.592 | 5514.865 | 4989.205 | 5091.754 | 5090.699 | 4979.048 | |
6 | 15,396.82 | 14,947.24 | 15,408.88 | 14,883.88 | 14,986.28 | 14,985.22 | 14,873.74 | |
7 | 450,75.86 | 44,626.9 | 45,087.91 | 44,563.58 | 44,665.81 | 44,664.76 | 44,553.44 | |
8 | 134,109.9 | 133,661.6 | 134,122 | 133,598.3 | 133,700.4 | 133,699.3 | 133,588.2 | |
9 | 401,209.2 | 400,761.4 | 401,221.2 | 400,698.2 | 400,800.1 | 400,799 | 400,688.1 | |
10 | 1,202,504 | 1,202,057 | 1,202,516 | 1,201,993 | 1,202,095 | 1,202,094 | 1,201,983 |
Station Order | BIC of the Daily API | |||||||
---|---|---|---|---|---|---|---|---|
KL | Klang | Cheras | PJ | Banting | ShA | KS | ||
Observed API | 1 | 644.1749 | 352.2063 | 696.8885 | 240.6926 | 448.2603 | 462.1786 | 178.31 |
2 | 663.4538 | 332.7529 | 698.4464 | 268.6937 | 458.9507 | 483.6555 | 239.9841 | |
3 | 827.6056 | 491.2444 | 828.3058 | 466.8144 | 636.4306 | 660.246 | 426.0242 | |
4 | 1548.058 | 1239.745 | 1539.657 | 1224.658 | 1381.723 | 1396.332 | 1184.039 | |
5 | 3778.04 | 3509.026 | 3762.801 | 3494.378 | 3635.01 | 3661.811 | 3453.944 | |
6 | 10,569.55 | 10,313.98 | 10,543.99 | 10,299.72 | 10,432.01 | 10,458.66 | 10,259.47 | |
7 | 30,942.22 | 30,725.85 | 30,933.25 | 30,711.95 | 30,841.64 | 30,868.32 | 30,670.89 | |
8 | 92,158.24 | 91,958.35 | 92,144.79 | 91,944.8 | 92,065.15 | 92,091.46 | 91,903.93 | |
9 | 275,833.6 | 275,651.6 | 275,812.2 | 275,639.5 | 275,750.7 | 275,770.9 | 275,598.9 | |
10 | 826,893.8 | 826,731.7 | 826,874 | 826,720 | 826,820.8 | 826,841.2 | 826,679.5 | |
Simulated API | 1 | 14,485 | 7672.153 | 15,679.53 | 5112.906 | 9859.986 | 10,138.47 | 3322.83 |
2 | 13,159.61 | 5918.581 | 14,025.42 | 4393.635 | 8522.06 | 9274.675 | 2777.936 | |
3 | 10,908.42 | 3692.711 | 11,496.32 | 2942.723 | 5883.349 | 7334.391 | 1802.015 | |
4 | 11,288.58 | 4726.253 | 11,762.87 | 3978.854 | 6792.007 | 8114.475 | 2890.721 | |
5 | 13,952.63 | 7960.398 | 14,301.05 | 7251.089 | 9894.717 | 11,170.95 | 6185.456 | |
6 | 23,201.69 | 17,810.61 | 23,490.54 | 17,118.36 | 19,639.26 | 20,878.37 | 16,065.42 | |
7 | 52,314.26 | 47,436.42 | 52,518.98 | 46,752.08 | 49,142.73 | 50,353.15 | 45,730.4 | |
8 | 140,860 | 136,418.4 | 141,000.5 | 135,751 | 138,021.5 | 139,172.8 | 134,752.5 | |
9 | 407,473.1 | 403,455.3 | 407,485.1 | 402,795 | 404,986.7 | 406,044.4 | 401,849.6 | |
10 | 1,208,346 | 1,204,661 | 1,208,368 | 1,204,055 | 1,206,154 | 1,207,143 | 1,203,142 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Alyousifi, Y.; Ibrahim, K.; Othamn, M.; Zin, W.Z.W.; Vergne, N.; Al-Yaari, A. Bayesian Information Criterion for Fitting the Optimum Order of Markov Chain Models: Methodology and Application to Air Pollution Data. Mathematics 2022, 10, 2280. https://doi.org/10.3390/math10132280
Alyousifi Y, Ibrahim K, Othamn M, Zin WZW, Vergne N, Al-Yaari A. Bayesian Information Criterion for Fitting the Optimum Order of Markov Chain Models: Methodology and Application to Air Pollution Data. Mathematics. 2022; 10(13):2280. https://doi.org/10.3390/math10132280
Chicago/Turabian StyleAlyousifi, Yousif, Kamarulzaman Ibrahim, Mahmod Othamn, Wan Zawiah Wan Zin, Nicolas Vergne, and Abdullah Al-Yaari. 2022. "Bayesian Information Criterion for Fitting the Optimum Order of Markov Chain Models: Methodology and Application to Air Pollution Data" Mathematics 10, no. 13: 2280. https://doi.org/10.3390/math10132280
APA StyleAlyousifi, Y., Ibrahim, K., Othamn, M., Zin, W. Z. W., Vergne, N., & Al-Yaari, A. (2022). Bayesian Information Criterion for Fitting the Optimum Order of Markov Chain Models: Methodology and Application to Air Pollution Data. Mathematics, 10(13), 2280. https://doi.org/10.3390/math10132280