Analyzing the Performance Criteria of ARMA Model for Air Quality Forecasting in Jakarta

Christina ALEXANDRA; Tiffany TANTRI; Fergyanto E. GUNAWAN

doi:10.14505//jemt.10.7(39).16

Christina ALEXANDRA Computer Science Department BINUS Graduate Program – Master of Computer Science Bina Nusantara University, Indonesia
Tiffany TANTRI Computer Science Department BINUS Graduate Program – Master of Computer Science Bina Nusantara University, Indonesia
Fergyanto E. GUNAWAN Industrial Engineering Department BINUS Graduate Program – Master of Industrial Engineering Bina Nusantara University, Indonesia

DOI: https://doi.org/10.14505//jemt.10.7(39).16

Abstract

Air pollution has been receiving global attention because of its effects on human health and the environment. The ability to predict the level of a pollutant concentration is important for various purposes such as for prevention and mitigation. Some papers had implemented various forecasting techniques to predict air pollution in Mexico, Spain, Malaysia. This research intends to investigate the forecasting accuracy of some air pollutants (PM10, SO2, CO, O3, and NO2) in Jakarta, Indonesia. The data are acquired daily for five measurement sites across the city, namely, Bundaran HI, Kelapa Gading, Jagakarsa, Lubang Buaya, and Kebon Jeruk during January-October of 2018. The data are fitted with Auto-Regressive Moving-Average (ARMA) models. The best model is obtained by comparing the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) with the relative error. The best and the most accurate model is the model with the lowest value of AIC or BIC.

References

[1] Abdulkareem, S. A. et al. 2019. Bayesian networks for spatial learning: a workflow on using limited survey data for intelligent learning in spatial agent-based models. Geoinformatica. Springer, 23(2): 243–268. DOI:https://doi.org/10.1007/s10707-019-00347-0
[2] Aho, K., Derryberry, D. and Peterson, T. 2014. Model selection for ecologists: the worldviews of AIC and BIC. Ecology. Wiley Online Library, 95(3): 631–636. DOI:https://doi.org/10.1890/13-1452.1
[3] Brewer, M. J., Butler, A. and Cooksley, S. L. 2016. The relative performance of AIC, AICC and BIC in the presence of unobserved heterogeneity. Methods in Ecology and Evolution. Wiley Online Library, 7(6): 679–692. DOI:https://doi.org/10.1111/2041-210X.12541
[4] Bulteel, K. et al. 2013. CHull as an alternative to AIC and BIC in the context of mixtures of factor analyzers. Behavior Research Methods. Springer, 45(3): 782–791. DOI:https://doi.org/10.3758/s13428-012-0293-y
[5] Contreras-Ochando, L. and Ferri, C. 2016. airVLC: An application for visualizing wind-sensitive interpolation of urban air pollution forecasts. Data Mining Workshops (ICDMW), 2016 IEEE 16th International Conference on, pp. 1296–1299. DOI: 10.1109/ICDMW.2016.0188
[6] Cortina--Januchs, M. G. et al. 2015. Development of a model for forecasting of PM10 concentrations in Salamanca, Mexico. Atmospheric Pollution Research. Elsevier, 6(4): 626–634. DOI:https://doi.org/10.5094/APR.2015.071
[7] Dziak, J. J. et al. 2019. Sensitivity and specificity of information criteria. bioRxiv. Cold Spring Harbor Laboratory, p. 449751. DOI: https://doi.org/10.1101/449751
[8] Emiliano, P. C., Vivanco, M. J. F. and De Menezes, F. S. 2014. Information criteria: How do they behave in different models?. Computational Statistics & Data Analysis, 69: 141–153. DOI:https://doi.org/10.1016/j.csda.2013.07.032
[9] Hafen, R. P. et al. 2014. Joint seasonal ARMA approach for modeling of load forecast errors in planning studies. in 2014 IEEE PES T&D Conference and Exposition, pp. 1–5. DOI:https://doi.org/10.1109/TDC.2014.6863150
[10] Lumley, T. and Scott, A. 2015. AIC and BIC for modeling with complex survey data. Journal of Survey Statistics and Methodology, 3(1): 1–18. DOI:https://doi.org/10.1093/jssam/smu021
[11] Medel, C. A. and Salgado, S. C. 2013. Does the BIC Estimate and Forecast Better than the AIC? Revista de Análisis Económico--Economic Analysis Review, 28(1): 47–64. DOI:https://doi.org/10.4067/s0718-88702013000100003
[12] Rahman, N. H. A. et al. 2015. Artificial neural networks and fuzzy time series forecasting: an application to air quality. Quality & Quantity, 49(6): 2633–2647. DOI:https://doi.org/10.1007/s11135-014-0132-6
[13] Schwarz, G. 1978. Estimating the dimension of a model. Annals of Statistics, 6 (2): 461--464.
[14] Wang, M. et al. 2016. A comparison of approaches to stepwise regression on variables sensitivities in building simulation and analysis. Energy and Buildings, 127: 313–326. DOI:https://doi.org/10.1016/j.enbuild.2016.05.065
[15] Wu, H., Cheung, S. F. and Leung, S. O. 2019. Simple use of BIC to Assess Model Selection Uncertainty: An Illustration using Mediation and Moderation Models. Multivariate behavioral research. Taylor & Francis, pp. 1–16. DOI:https://doi.org/10.1080/00273171.2019.1574546