1917

Air Quality Assessment by Statistical Learning-Based Regularization

PM10 can be stated as a particulate matter smaller than 10 micrometer and it can be suspended in the air. The incremental concentration of PM10 affects both human and environment drastically. In this study, an air quality assessment by exhibiting the potential relationships among the secondary indicators and PM10 has been focused. For the analyses, statistical learning-based regularization procedures such as Ridge, the Lasso and Elastic-net algorithms have been practiced. In particular, use of Elastic-net algorithm in predicting PM10 concentration includes a novelty. As a result of the computational studies, it has been recorded that all the models showed high accuracy capacities. However, the elastic-net model outperforms the other models both accuracy and robustness (stability). Considering the errormeasurements (MSE and MAPE), the best numerical results have been provided by the Elastic-net model. Use of machine learning-based regularization algorithms in environmental problems can provide accurate model structures as well as generality and transparency.

İstatistiksel Öğrenmeye Dayalı Düzenlemeyle Hava Kalitesinin Değerlendirilmesi

PM10, 10 mikrometreden daha küçük boyutta, havada askıda kalma özelliğine sahip parçacık madde olarak tanımlanabilir. PM10’un çok yüksek konsantrasyonları insan ve çevreyi şiddetli biçimde etkiler. Bu çalışmada, hava kalitesinin değerlendirilmesi amacıyla, ikincil parametreler ile PM10 arasındaki ilişkilerin ortaya çıkarılmasına odaklanılmıştır. Analizler için istatistiksel öğrenmeye dayalı düzenleme yöntemleri olan Ridge, Lasso ve Elastic-net yordamlarından yararlanılmıştır. Özellikle Elastic-net yordamının PM10 tahmininde kullanımı yenilik taşımaktadır. Hesaplamaların sonucu olarak, bütün modellerin yüksek kestirim kapasitesine sahip oldukları kaydedilmiştir. Bununla birlikte, gerek kestirim başarısı ve gerekse de model gürbüzlüğü (duraylılığı) bakımından Elastic-net modeli diğer yöntemlerle karşılaştırıldığında daha başarılı sonuçlar vermektedir. Model hata ölçümleri (MSE ve MAPE) temel alındığında, en iyi sayısal sonuçlar Elastic-net modeliyle elde edilmiştir. Makine öğrenmesine dayalı düzenleme yordamlarının çevresel problemlerin değerlendirilmesi amacıyla kullanımı başarılı, genelleştirilmiş ve şeffaf model yapılarının oluşturulmasını sağlayabilecektir.

PDF

___

1. Mallik, C., 2019. Anthropogenic Sources of Air Pollution, in Air Pollution: Sources, ed. Impacts and Controls, Saxena, P., Naik, V., CABI. New Delhi.

2. Radzka, E., Rymuza, K., 2019. The Effect of Meteorological Conditions on PM10 and PM2.5 Pollution of the Air. Rocznık Ochrona Srodowiska 21(1), 611-628.

3. Lai, L.W., 2016. Public Health Risks of Prolonged Fine Particle Events Associated with Stagnation and Air Quality Index Based on Fine Particle Matter with Diameter <2.5 mμ in the Kaoping Region of Taiwan. Int. J. of Biometeorology, 60(12), 1907-1917.

4. Nguyen, G.T.H., Shimadera, H., Uranishi, K., Matsuo, T., Kondo, A., Thepanondh, S., 2019. Numerical Assessment of PM2.5 and 0-3 Air Quality in Continental Southeast Asia: Baseline Simulation and Aerosol Direct Effects Investigation. Atmospheric Environment, 219, 117064.

5. Yatkin, S., Gerboles, M., Belis, C.A., Karagulian, F., Lagler, F., Barbiere, M., Borowlak, A., 2020. Representativeness of an Air Quality Monitoring Station for PM2.5 and Source Apportionment Over a Small Urban Domain. Atmospheric Pollution Research, 11(2), 225-233.7

6. Alvarez-Mendoza, C.I., Teodoro, A.C., Torres, N., Vivanco, V., 2019. Assessment of Remote Sensing Data to Model PM10 Estimation in Cities with a Low Number of Air Quality Stations: A Case of Study in Quito. Ecuador, Environments, 6(7), 85.

7. Petrowski, K., Bastianon, C.D., Buhrer, S., Brahler, E., 2019. Air Quality and Chronic Stress a Representative Study of Air Pollution (PM2.5, PM10) in Germany. J. Occupational and Environmental Medicine, 61(2), 144-147.

8. Yoon, H., 2019. Effects of Particulate Matter (PM10) on Tourism Sales Revenue: a Generalized Additive Modelling Approach. Tourism Management, 74, 358-369.

9. Akdi, Y., Okkaoglu, Y., Golveren, E., Yucel, M.E., 2020. Estimation and Forecasting of PM10 Air Pollution in Ankara Via Time Series and Harmonic Regressions. Int. J. Environmental Science and Technology, https://doi.org/10.1007/s13762-020-02705-0.

10. Draper, N.R., Smith, H., 1998. Applied Regression Analysis, Wiley, USA.

11. Saleh, A.K.M.E., Arashi, M., Kibria, B.M.G., 2019. Theory of Ridge Regression with Applications, John Wiley & Sons, USA.

12. Tutmez, B., 2018. Bauxite Quality Classification by Shrinkage Methods, Journal of Geochemical Exploration, 191, 22-27.

13. Zou, H., Hastie, T., 2005. Regularization and Variable Selection Via the Elastic Net. Journal of the Royal Statistical Society, Series B:301-320.

14. Megaritis, A.G., Fountoukis, C., Charalampidis, P.E., Pilinis, C., Pandis, S.N., 2013. Response of Fine Particulate Matter Concentrations to Changes Ofemissions and Temperature in Europe. Atmos. Chem. Phys.,13, 3423–3443.

15. James, G., Witten, D., Hastie, T., Tibshirani, R., 2013. An Introduction to Statistical Learning, Springer, New York.

16. Dorugade, A.V., 2014. New Ridge Parameters for Ridge Regression. Journal the Association of Arab Universities for Basic and Applied Sciences, 15(1), 94-99.

17. Hastie, T., Tibshirani, R., Wainwright, M., 2015. Statistical Learning with Sparsity, CRC Press, Boca Raton.

18. Kuhn, M., Johnson, K., 2013. Applied Predictive Modelling, Springer, New York.

19. Khan, M.H.R., Anamika, B., Tamanna, H., 2019. Stability Selection for Lasso, Ridge and Elastic net Implemented with AFT Models, Statistical Applications in Genetics and Molecular Biology, 18(5), 10.1515/sagmb-2017-0001.

20. ÇŞB., 2014. Eskişehir İli Temiz Hava Eylem Planı, THEP (2014-2019), Eskişehir. (in Turkish).

21. Tutmez, B., 2019. Multivariate Statistical Control of Air Quality. 2. International Mersin Symposium, Mersin, 370-381.

22. R Development Core Team, 2008. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria (ISBN 3-900051-07-0).

23.Friedman, J., Hastie, T., Tibshirani, R., 2010. Regularization Paths for Generalized Linear Models Via Coordinate Descent. J. Statistical Softwares, 33, 1–22.

24. Kuhn, M., 2008. Building Predictive Models in R Using the Caret Package. Journal of Statistical Software 28(5), 1-26.

25. Alpaydın, E., 2010. Introduction to Machine Learning, the MIT Press, Cambridge.