PM2.5_Time_Series_Analysis

ARIMAX and RNN+LSTM methods have been implemented for the prediction on PM2.5 (superfine particles in the air which are harmful to human health) by using the dataset “PRSA2017_Data_20130301-20170228.zip”. The dataset contains air pollutants data from 12 monitoring sites in China. The data was collected at regular 1-hour intervals from 01-Mar-2013 to 28-Feb-2017. Both the methods have been developed and their prediction performances have been evaluated and compared on each of the 12 sites. Both .py and Jupyter Notebook (.ipynb) files have been uploaded for convinience. For the program to work, the code file and datasets (.csv files) need to be in the same working directory.