Forecasting And Prediction Of Air Pollutants Concentrates Using Machine Learning Techniques: The Case Of India

Abstract Air quality index (AQI) is a number used by government agencies to communicate the public how polluted air currently. It based on several factors like SO2, NO2, O3, RSPM/PM10, and PM2.5. Several methods were developed in past various researchers/environmental for determination of AQI. Still, there no universally accepted method that exists, which appropriate all situations. We have prediction model confined standard classification or regression models. These models ignored co-relation between sub-models different time slots. The paper focusses refined inferring pollutants historical current meteorological datasets. Also, designed forecast AQI coming months, quarters years where emphasis improve its accuracy performance. algorithms are Pollution Geocodes Dataset (2016-2018), results calculated 196 cities India classifiers. Accuracy 94%-96% achieved from Linear Robust Regression, increases 97.92% after application KNN 97.91% SVM 97.47 5th epoch ANN. Decision Tree Classifier has given best 99.7%, 0.02% Random Forest Classifier. Forecasting Moving Average Smoothing using R-ARIMA, offers daily values 45days monthly data next year.