A Comparison between Data Mining Techniques and Time Series Models for Forecasting Air Pollution in Baghdad

Abstract

Concern for the environment is an important priority in different countries, Environmental pollution is the most important source of environmental threat. Pollution levels in the water, air and land environment have reached serious limits, requiring researchers in various sciences to take care of researches that reduce, monitor and reduce their causes within the limits allowed. Air pollution is one of the main threats to environmental pollution, which has a direct impact on human life. It is due to the increase in the temperature of the earth and the depletion of the ozone layer due to the dangerous emissions of gases directly into the atmosphere, mainly NO2 and SO2 . This paper aims to compare the Box & Jenkins method that include Auto Regressive Integrated Moving Average models (ARIMA), Vector Auto regressive models (VAR) and data mining techniques that include Artificial Neural Network (ANN). The data gathered from Baghdad city for the period 2015-2017 with 157 observations. This paper showed the superiority ANN models compared to VAR and ARIMA models, as well as VAR models compared to ARIMA models.