Clustering the Concentrations of PM10 and O3: Application of Spatiotemporal Model–Based Clustering

Air pollution data are large-scale datasets that can be analyzed in low scales by clustering to recognize the pattern of and have simpler more comprehensible interpretations. So, this study aims cluster days year 2017 according hourly O3 PM10 amounts collected from four stations Tabriz using spatiotemporal mixture model–based (STMC). Besides, with temporal dimension (TMC) without considering dimensions (MC) were utilized compare STMC. To evaluate efficiency these three models, obtain optimal number clusters each model, BIC ICL criteria used. According ICL, STMC outperforms TMC MC. Three for selected as fit models. Regarding PM10, average concentration was highest 4. O3, all summer 3, highest. Cluster 2 had lowest a high difference 1 its temperature lowest. Autumn make up about 84% cluster. The polluted clean into separate groups observing effect meteorological factors on amount clearly prove model. Results showed air increases both dimensions.