Predictions of PM 2.5 using air pollutants and meteorological factors with COVID-19 cases in Malaysia and Indonesia: a comparative study using feature selection and robust regression.

Autor: Salsabila NB; Department of Environmental and Occupational Health, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Malaysia., Jalaludin J; Department of Environmental and Occupational Health, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Malaysia., Suhaimi NF; Department of Environmental and Occupational Health, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Malaysia., Wan Mansor WN; Faculty of Ocean Engineering Technology, Universiti Malaysia Terengganu, Kuala Nerus, Malaysia., Sumantri A; Study Program of Public Health, Health Science Faculty, State Islamic University (UIN), Syarif Hidayatullah Jakarta, Indonesia.
Jazyk: angličtina
Zdroj: International journal of environmental health research [Int J Environ Health Res] 2024 Aug 13, pp. 1-22. Date of Electronic Publication: 2024 Aug 13.
DOI: 10.1080/09603123.2024.2390479
Abstrakt: The study examines the relationship between air quality, meteorological factors, and COVID-19 cases in Cheras, Kuala Lumpur, and Kelapa Gading, North Jakarta. Analyzing data from 2020 and 2021, the research found notable correlations: COVID-19 cases in Cheras were positively associated with relative humidity (RH) and carbon monoxide (CO) but negatively with ozone (O₃) and RH in different years. In Kelapa Gading, COVID-19 cases were positively correlated with pollutants like sulfur dioxide (SO₂) and CO, while ambient temperature (AT) showed a negative correlation. The enforcement of social restrictions notably reduced air pollution, affecting COVID-19 spread. Predictive models for PM 2.5 levels using robust regression techniques showed strong performance in Kuala Lumpur (R² > 0.9) but exhibited overfitting tendencies in Jakarta, suggesting the need for a longer study period for more accurate results.
Databáze: MEDLINE