Abdullah, Muhammad Hisyam Lee and Javedani, Hossein and Suhartono, Suhartono (2010) An evaluation of some classical methods for forecasting electricity usage on specific problem. In: The Regional Conference on Statistical Sciences (RCSS'10), 13-14 June 2010, New Pacific Hotel, Kota Bharu, Kelantan.
Full text not available from this repository.
This empirical paper compares the accuracy of five univariate methods for quarterly electricity demand forecasting, namely naïve, regression, decomposition additive and multiplicative, exponential smoothing, and Box-Jenkins methods. Data generated from 66 quarters electricity usage in Washington power supply are used as case study. The data are divided into two parts, namely in-sample and out-sample for parameter estimations and forecasting evaluation respectively. The results show that when the data contain some outliers, ARIMA model may give the unacceptable results. In contrast exponential smoothing methods are suitable in this condition because it gives more weight to the most recently observation. In addition, the performance evaluation shows that an exponential smoothing method yields more accurate forecast than other methods.
|Item Type:||Conference or Workshop Item (Paper)|
|Uncontrolled Keywords:||seasonal data, naïve, regression, decomposition, exponential smoothing, box-jenkins|
|Deposited By:||Liza Porijo|
|Deposited On:||13 Jun 2012 01:05|
|Last Modified:||13 Jun 2012 01:05|
Repository Staff Only: item control page