Classification of Air Quality Index Using Autoregressive Integrated Moving Average (ARIMA) and Artificial Neural Networks Models

One of the negative effects of urbanization and population increase in developing nations is air pollution. As a result, the Air Quality Index (AQI) is assessed in a descriptive system that serves as a medium for informing the public about the danger of pollution. Thus, a deep learning system automates the assessment of air contaminants using time series data. Over the years, one of the most often used linear models in time series forecasting has been the Autoregressive Integrated Moving Average (ARIMA). Artificial neural networks (ANNs) have been the subject of recent study into forecasting, and these studies indicate that ANNs may prove to be a viable substitute for more conventional linear methods. In order to take advantage on the distinct advantages of ARIMA and ANN models in linear and nonlinear modelling, a hybrid methodology combining both models are presented in this study. This study made use of the data from the UCI Machine Learning Repository 2017. The primary objective of the Long Short-Term Model (LSTM), which is used for classification, is to improve accuracy in the vicinity of the test point. Based on the findings, the suggested ARIMA-ANN model outperforms the current model in terms of accuracy. The combined model can be an efficient means of enhancing the predicting accuracy attained by either of the models employed separately, according to the study results with the given data sets.