Spatio-Temporal Clustering Analysis for Air Pollution Particulate Matter (PM2.5) Using a Deep Learning Model

Abstract

Fine particulate issue (PM2.5) is normal air contamination and has antagonistic well-being impacts globally, particularly in the quickly industrial nation such as Taiwan because of massive air contamination. The PM2.5 pollution changes with existence separation and is overwhelmed by the area’s limitation inferable from the distinction’s uniqueness in topographical conditions including geology and meteorology, and the trademark’s gadget normal for urbanization and industrialization. To portray these boundaries and components, contention, and mechanics, the five-years PM2.5 contamination examples of Newport area’s duty in eastern Taiwan with high-goal senior high goal were explored. This resolution was found using the linear assignment to build the clustering model with a convolution autoencoder for Spatio-temporal analysis for air pollution particulate matter PM2.5. In all the above models fully connected model is a better result performance model.