Forecasting Electricity Generation Capacity in Malaysia: An Auto Regressive Integrated Moving Average Approach

Haiges, Rina; Wang, Y.D.; Ghoshray, A.; Roskilly, A.P.

Abstract

It is imperative for Malaysia to have a clear understanding of the future performance of its power sector with emphasis on the total installed capacity variable as this is integral to support the nation's capacity succession planning over an intermediate to long term period in order to sustain the economy. This paper aims to deploy the Auto Regressive Integrated Moving Average (ARIMA) approach to fit the 40 years forecast up to 2053 by assessing 40 years of past data from 1973 until 2013. The different models will be evaluated using the Schwarz Bayesian Criterion (SBC). Validation was performed by comparison of forecast and actual data based on a five-year holdback period. Accuracy measures applied were the Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). In this assessment, ARIMA(0,2) demonstrated a better forecast in terms of accuracy during the holdback period. However, the Diebold-Mariano (DM) test didn’t detect any differences between the ARIMA(1,0) and ARIMA(0,2) forecasts. Application of the forecast results was demonstrated as well.

• Presentation Conference Type: Conference Paper (Published)
• Conference Name: The 8th International Conference on Applied Energy – ICAE2016
• Start Date: Oct 8, 2016
• End Date: Oct 11, 2016
• Online Publication Date: Jun 1, 2017
• Publication Date: 2017-05
• Deposit Date: Nov 5, 2019
• Publisher: Elsevier
• Volume: 105
• Pages: 3471-3478
• Series Title: Energy Procedia
• Series ISSN: 1876-6102
• DOI: https://doi.org/10.1016/j.egypro.2017.03.795
• Keywords: ARIMA; electricity; forecasting; installed capacity; power sector; Malaysia
• Public URL: https://durham-repository.worktribe.com/output/1278781