| Abstrakt: |
In this research, we propose an integrated approach to retrieve the information of air quality monitoring data for air quality assessment. Used data is the hourly concentrations of PM10, PM2.5, SO2, NO, NO2, O3, and CO measured by an ambient air quality monitoring station in Hanoi, Vietnam, in the period of 2010–2018. First, the dataset was preprocessed to remove outliers by a simple removing method. Then, the missing values were filled up using a deep learning technique. After that, the preprocessed data went through the process of analysis to infer the probability distributions. After the probability density functions (pdf) of the hourly concentrations of these air pollutants in the year were determined, the information of air quality data in each year including the annual mean and median, the maximum concentration, and the number of exceedances of hourly thresholds for the period of 2010–2018 was obtained. High level of PM2.5, exceeding its national annual standard, was observed implying the potential of health risk to the people. Trend and temporal variations of these air pollutants were also analyzed. It is found out that the effect of direct emission from local traffic activity on the ambient levels of SO2 and PM2.5 is not dominant. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink