Prediction models using outdoor environmental data for real-time PM10 concentrations in daycare centers, kindergartens, and elementary schools

Abstract

Abstract Children spend a considerable amount of time in daycare centers, kindergartens, and elementary schools. Poor indoor air quality (IAQ) in the educational facilities can affect the health of the children and impair their academic performance. The prediction of real-time PM10 concentration could be useful to intervene the problem of poor IAQ. This study developed models to predict real-time indoor PM10 concentration in the daycare centers, kindergartens, and elementary schools using outdoor environmental data. Indoor PM10 concentrations were measured in 54 daycare centers, 12 kindergartens, and 21 elementary schools in Seoul, South Korea, using a real-time monitor (AirGuard K) over a period of one year. Multiple linear regression models were used to predict real-time indoor PM10 concentration in these educational facilities using outdoor PM10 and meteorological data as input variable. Four formations (original, ratio of indoor-to-outdoor, root-transformation, and log-transformation) for dependent variable were compared to determine the best performance of the model. A 10-fold cross-validation method was used to evaluate the accuracy of the prediction models. Daycare centers showed the highest indoor PM10 concentration. Root-transformed models with high accuracy were developed to predict the real-time indoor PM10 concentration in educational facilities every 10\u202fmin. The R2 of the prediction models were 0.64 in the daycare centers, 0.45 in the kindergartens, and 0.43 in the elementary schools. The 24\u202fh profile of the predicted indoor PM10 was similar to the measured PM10 concentration. The prediction models could provide real-time PM10 levels in educational facilities without direct indoor measurement and observation.