印尼蝦輸日之預測分析

Forecasting Analysis of Indonesia’s Shrimp Export to Japan

羅克曼(Rhochmad Wahyu Illahi)

指導教授 ： 陳淑恩 ； -

國立屏東科技大學/國際學院/農企業管理國際碩士學位學程/碩士(2019年)

https://doi.org/10.6346/THE.NPUST.AM.020.2019.D04

摘要

印度尼西亞是世界上最大的蝦類生產國，主要出口到美國和日本，佔蝦出口的三分之一以上。蝦捕獲結果和出口在不同季節和不同年份波動，故在經營管理和農業政策上採用準確方法來預測出口，是非常重要。本研究的研究目的是分析預測印度尼西亞蝦出口到日本的最佳模型。
印度尼西亞蝦出口數據是1989-2017年的時間序列。本研究利用移動平均法，單指數平滑法，最小平方趨勢法，二次趨勢法和Box-Jenkins（ARIMA）模型，以1989-2017年期間印度尼西亞蝦出口到日本資料，來預測2018年出口。然後，預測誤差分析是利用MAPE（平均絕對百分比誤差），MAD（平均絕對偏差）和MSE（均方誤差）分析。本研究數據分析是使用Minitab 第17版和Microsoft Excel。
本研究預測2018年印度尼西亞蝦出口到日本的結果，以3期之移動平均模型預測值為27,080,419公斤；以單指數平滑模型平滑係數⍺分別為0.2、0.5、0.9，預測值分別為30,942,641公斤、28,444,131公斤、和30,191,341公斤。最小平方趨勢模型預測為27,101,974.88公斤。二次趨勢趨勢模型預測為19,089,500公斤，ARIMA模型預測（0,1,1）為24,213,800公斤。
最後，因為ARIMA（0,1,1）模型在MAPE、MAD、MSE皆具有最小的預測誤差值，表示其比其他預測方法更能準確地預測，故本研究選擇ARIMA（0,1,1）為最佳預測模型。

關鍵字

蝦出口，預測，印度尼西亞，ARIMA ； 蝦出口 ； 預測 ； 印度尼西亞 ； ARIMA

並列摘要

Indonesia, the largest shrimp producer in the world, exports shrimp mainly to the US and Japan, accounting for more than one third of the shrimp export. The shrimp catch results and exports fluctuate in different seasons and different years. It is important to adopt an accuracy method to forecast export as a guideline for management strategy and agricultural policy. The purpose of this study was to analyze the best model to forecast Indonesia’s shrimp export to Japan.
Data of Indonesia’s Shrimp exports are time series during 1989-2017. Moving Averages, Single Exponential Smoothing, Least Squares Trend, Quadratic Trend, and Box-Jenkins (ARIMA) models were used to predict the Indonesia’s shrimp export to Japan during 1989-2017 and to forecast export in 2018. Then, the forecast error analysis was analyzed by MAPE (Mean Absolute Percentage Error), MAD (Mean Absolute Deviation), and MSE (Mean Square Error). Minitab 17 application and Microsoft Excel were used for data analysis.

By 3-period Moving Average model, the Indonesia's shrimp export to Japan in 2018 is forecast as 27,080,419 kg. By Single Exponential Smoothing Model with smoothing coefficient ⍺ equaling 0.2, 0.5, and 0.9, shrimp exports are forecast as 30,942,641 kg, 28,444,131 kg, and 30,191,341 kg, respectively. Forecast with Least Squares Trend model is 27,101,974.88 kg. Forecasts of shrimp exports are 19,089,500 kg and 24,213,800 kg respectively through Quadratic Trend Model and ARIMA model (0, 1, 1).
As a result, ARIMA (0, 1, 1) was selected as the best model as forecasting model because it has the smallest forecasting errors evaluated by MAPE, MAD, and MSE and is able to predict more accurately than other forecasting models.

並列關鍵字

Shrimp Export, Forecast, Indonesia, ARIMA ； Shrimp Export ； Forecast ； Indonesia ； ARIMA

參考文獻

Ajeng, S. 2011. Demand Forecasting for Planning the Procurement of Durian Fruit Supplies in Durian Harum Bintaro Place, Jakarta. Thesis. University of Islam Negeri Syarif Hidyatullah. Jakarta.

Arsyad, L. 1995. Business Forecasting. BPFE. Yogyakarta.

Assauri, S. 1984. Techniques & Method of Forecasting: Applications in Economic & Business. Edition 1. Publisher Institution Faculty of Economic University of Indonesia. Jakarta

Dian. 2006. Classification and Morphology Vannamei Shrimp (Litopenaeus vannamei)

Directorate General of National Export Development. 2014. Export News: Fish and Fish Products. Directorate General of National Export Development, Ministry of Commerce. Jakarta