Far East Journal of Experimental and Theoretical Artificial Intelligence
Volume 3, Issue 2 (Special Issue on 13th Conference on Artificial Intelligence and Applications (TAAI 2008), Pages 125 - 141
(May 2009)
|
|
EXPONENTIAL SMOOTHING FORECASTING MODELS FOR FISHERY STATISTICS
Hsin-Wei Wang (Taiwan), Tun-Wen Pai (Taiwan) and Chia-Han Chu (Taiwan)
|
Abstract:
An exponential smoothing forecasting model is built to analyze and forecast the amount and unit price of export and import fishery products. Based on the statistics of historical data, the amount and unit price of each fish product in the coming year can be predicted, which provides useful reference information on decision-making for fishing trades and related businesses. Four different models of the moving average, horizontal seasonal smoothing forecasting, additive seasonal smoothing forecasting, and multiplicative seasonal smoothing forecasting models are designed as fundamental archetypes in the proposed forecasting system. The proposed models reflect periodic and trend properties from historical time-series data, and rapidly adjust and improve predicted precision according to the updated values. From the experimental data, the amount and unit price of most import and export fishery products in Taiwan are demonstrated with seasonal changing factors and periodic properties, and the prediction error rates for quantity and price are shown and discussed in this paper. |
| Keywords and phrases: exponential smoothing forecasting, time series, periodicity, seasonal ratio, fishery statistics. |
| |
| Number of Downloads: 146 | Number of Views: 376 |
|
