Estimation of the Visibility in Seoul, South Korea, Based on Particulate Matter and Weather Data, Using Machine-learning Algorithm

Visibility is an important indicator of air quality and any consequent meteorological climate change. Therefore, visibility in Seoul, which the most polluted city South Korea, was estimated using machine learning (ML) algorithms based on (temperature, relative humidity, precipitation) particulate matter (PM10 PM2.5) data acquired from automatic weather station, compared with observed visibility. Meteorological data, at 1-h intervals between 2018 2020, were used. Through validation each ML algorithm, extreme gradient boosting (XGB) algorithm found to be suitable for estimations (bias = 0 km, root mean square error (RMSE) 0.08 r 1 training set). Among used XGB importance PM2.5 humidity variables high (51% 19%, respectively), whereas precipitation wind speed had low (approximately 1%). The estimation accuracy test dataset good –0.11 RMSE 2.08 0.94); higher dry season –0.06 1.79 0.96) than rainy –0.17 2.34 0.91). results this study indicated a correlation previous studies. proposed method promotes accurate areas poor visibility, thus, it can assess public health quality.