Forecasting export of selected timber products from Peninsular Malaysia using time series analysis
The export of timber products from Peninsular Malaysia is an important economic trade for the country. With the changes in world’s economies, it highlights the need to apply the forecasting methods in anticipating the trend in the export of timber products from Peninsular Malaysia. This study was do...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2011
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/27863/1/FH%202011%2021R.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The export of timber products from Peninsular Malaysia is an important economic trade for the country. With the changes in world’s economies, it highlights the need to apply the forecasting methods in anticipating the trend in the export of timber products from Peninsular Malaysia. This study was done by analyzing 110 quarterly observations data (from March 1982 to June 2009) of sawntimber, mouldings and chipboard volume (m³) with four time series methods (the Seasonal Holt-Winters and ARAR algorithms as well as the ARMA and Seasonal ARIMA models). The quarterly observations data was taken from the Report on Timber Export Statistics (Peninsular Malaysia), published by the Malaysia Timber Industry Board (MTIB) Resource Centre. The data were divided into two portions where the first 100 quarterly observations (calibration data set or within-sample data) were used in the modelling process. The remaining ten quarterly observations (validation data set or out-of-sample data)were used to assess the forecasting abilities based on the measures of accuracy including mean absolute error (MAE), root mean square error (RMSE) and mean absolute percentage error (MAPE). MAPE was considered to be the decisive factors in measuring the accuracy of the forecasts as it presented different levels of model accuracy evaluations. Results have shown that the modelling process on the within-sample data in the export of sawntimber indicated the ARAR algorithm had produced the best forecast. From the assessments on the out-of-sample data, the forecasting abilities showed ARAR algorithm had the lowest MAPE at 17.27%. For a six quarters period into the future, the estimated exports of sawntimber range from 100,000 m3 to 700,000 m3 at 95% confidence intervals. For mouldings and chipboard, the modelling process showed that the Seasonal ARIMA (1, 0, 4) X (0, 1, 0)4 model produced the best forecasts. The assessments on the out-of-sample data for the Seasonal ARIMA (1, 0, 4) X (0, 1, 0)4 model showed the forecasting abilities with the lowest forecast errors where MAPE was at 18.83%. For the export in six quarters ahead, the forecasts are expected more than 150,000 m3 at 95% confidence level. The study concluded that the forecasts offer favorable amounts that based on the assumptions that all related events in the export for these timber products will not drastically change the forecasts. This study illustrates the anticipated trend in the export of the selected timber products from Peninsular Malaysia that both public and private sectors could utilize in their decision making of future planning in order to meet the export demand. |
|---|