An Assessment of Renewable Energy in Bangladesh through ARIMA, Holt’s, ARCH-GARCH Models

Authors

  • Fatema Tuz Jhohura
  • Md. Israt Rayhan

Keywords:

Autoregressive integrated moving average, Holt’s linear exponential smoothing, Autoregressive conditional heteroscedastic, Akaike information criteria, Schwartz Bayesian criteria

Abstract

Forecasting of the Renewable Energy plays a major role in optimal decision formula for government and industrial sector in Bangladesh. This research is based on time series modeling with special application to solar energy data for Dhaka city. Three families of time series models namely, the autoregressive integrated moving average models, Holt’s linear exponential smoothing, and the autoregressive conditional heteroscedastic (with their extensions to generalized autoregressive conditional heteroscedastic) models were fitted to the data. The goodness of fit is performed via the Akaike information criteria, Schwartz Bayesian criteria. It was established that the generalized autoregressive conditional heteroscedastic model was superior to the autoregressive integrated moving average model and Holt’s linear exponential smoothing because the data was characterized by changing mean and variance.

Downloads

Issue

Section

Articles