Air pollution forecasting using multiple time series approach

Abstract

Air pollution forecasting helps to take precautionary measures in order to maintain public health. Time series algorithms and software such as Prophet package are used to forecast the air quality level in environment. A univariate data makes it easy to apply time series algorithms such as auto regressive integrated moving average (ARIMA) and Naive Bayes. The results of Naive Bayes, Prophet package and ARIMA are represented in 2-D plane just to better understand and analyze the output. Comparison of all the above said algorithms and software is done to come up with the best approach.