Forecasting Particulate Matter Emissions Using Time Series Models
2023
Suresh, S. | Sindhumol, M. R. | Ramadurai, M. | Kalvinithi, D. | Sangeetha, M.
Environmental pollution is a serious concern nowadays with its disastrous impact on living organisms. In several types of pollution, Air pollution takes on a crucial role by directly affecting the respiratory system and causing fatal diseases in humans. Air pollution is a mixture of gaseous and particulate matter interweaved by different sources and emanating into the atmosphere. In particular, particle pollutants are critical in growing air pollution in India’s main cities. Forecasting the particulate matter could mitigate the complications caused by it. The employment of a model to predict future values based on previously observed values is known as time series forecasting. In this paper, the PM2.5 pollutant emission data recorded at the Kodungaiyur region of Chennai city were forecasted using three-time series models. The standard ARIMA model is compared with the deep learning-based LSTM model and Facebook’s developed Prophet algorithm. This comparison helps to identify an appropriate forecasting model for PM2.5 pollutant emission. The Root Mean Squared Error (RMSE) acquired from experimental findings is used to compare model performances.
显示更多 [+] 显示较少 [-]