鵝價預測之動態分析

A Dynamic Analysis of Geese Predicted Prices

謝志忠(Chieh-Chung Hsieh)

《管理實務與理論研究》 2卷2期 (2008/06) Pp. 139-152

https://doi.org/10.29916/JMPP.200806.0008

摘要

本研究乃分析1998-2006年共459週的週資料，欲瞭解養鵝產業的小鵝價格和肉額價格的統計特性及其價格預測。爲了讓分析結果能夠普及至一般鵝農，使之能夠瞭解閱讀參閱，研究方法採用一般時間數列分析方法，但仍不失其模型的準確性。經過分析後，獲得如下的統計結果：1.小鵝平均價格在52.34元／隻，肉鵝平均價格在33.33元／公斤。兩者相關係數只有0.22。2.經過Augumented Dickey-Fuller單根檢定後，兩個時間數列呈現穩定時間數列的特性。3.其估計的迴歸預測模型如下：a.小鵝價格迴歸模型：SGP=0.9564190856(上标 *)SGP(-1) + 0.5227434296(上标 *)D(SGP(-1)) + 0.09164880255(上标 *)D(SGP(-2)) +0.05546035066(上标 *)D(SGP(-3)) + 0.1797128706(上标 *)D(SGP(-4)) + 2.288827474 b. 肉鵝價格迴歸模型：GP=0.9580367044(上标 *)GP(-1) + 0.5677031589(上标 *)D(GP(-1)) + 1.392161044。由分析結果顯示：1.現代化的農民是否要每期均辛勤飼養肉鵝有待商確。蓋肉鵝價格波動幅度满大的，在價格下跌時，是否應將勞動力移作它用，例如作農場的衛生消毒工作，爲下一期健康鵝隻的飼養作好準備，此均有賴農場主看懂價格變動歷史圖型及分析結果進行最佳決策。2.迴歸預測模型顯示小鵝價格受前面四期的小鵝價格的影響，而肉鵝的平均價格則受前一天價格的影響。因此鵝農在作價格預測時應考慮最近幾日的價格波動。

關鍵字

鵝 ； 價格預測 ； 動態模型

並列摘要

An analysis of the price fluctuations of goose industry during the 459 week period commencing in 1998 and ending in 2006 was carried out in order to unveil the statistical characteristics of goose prices and to forecast the prices. The time series analysis method was considered appropriate because it is easy for goose farmers to understand and ensures predication accuracy. Several important findings are noted in the present study: 1. The average sale price per goose was NT$52.34 and NT$ 33.33 per kilogram. A correlation coefficient was r=0.22. 2. The time series becomes relatively stable after the first order difference by the augmented Dickey-Fuller test. 3. The models are proposed below: a. The regression model for estimation of goose price is: SGP=0.9564190856(superscript *)SGP(-1) + 0.5227434296(superscript *)D(SGP(-1)) + 0.09164880255(superscript *)D(SGP(-2)) + 0.05546035066(superscript *)D(SGP(-3)) + 0.1797128706(superscript *)D(SGP(-4)) + 2.288827474. b. The regression model for estimation of goose meat price is: GP=0.9580367044(superscript *)GP(-1) + 0.5677031589(superscript *)D(GP(-1)) + 1.392161044. In summary, the results of the present study demonstrate that: 1. It remains to be seen whether farmers should still be committed to the goose industry. Its history shows big fluctuations in its prices. Goose farmers should be aware that price fluctuations could have a severe impact on the movement of the labor force as well as the quality of their decision-making. For example, when the price falls, the labor force should be allowed to shrink or move to work in other areas such as the cleaning and disinfection of farm premises for preparation of the next flock. 2. The goose price is likely to be affected by the prices over the previous four periods whereas the goose meat price might be influenced by the price in the previous day. Goose farmers should take into account the recent movement in both the goose price and meat price while making prediction.

並列關鍵字

gooses ； price forecast ； dynamic model

參考文獻

李宗儒、鄭卉方(2000)。應用灰色理論預測物之價格－以紅豆爲例。農林學報。49(2)，83-92。

林茂文(1992)。時間數列分析與預測。臺北:華泰書局。

林銘昌(2002)。臺灣芒果批發價格預測之比較研究。國立臺灣大學農業經濟研究所。

吳柏林(1995)。時間數列分析導論。臺北:華泰書局。

唐淑娟(2001)。臺彎地區鳳梨零售價格預測之研究－灰預測、類神經網路與預測組合之應用。國立屏東科技大學農企業管理研究所。