- 學位論文
高雄果菜運銷公司高麗菜批發價格預測之研究 -蒐集私有資料的價值
Forecasting Weekly Cabbage Price in Kaohsiung, Taiwan: The Value of Private Data
摘要
本研究目的是探討私有資料在供需市場的價值。資料包含公開及私有資料。私有資料是經由高麗菜中盤商所蒐集，用以預測下周高雄果菜運銷公司高麗菜批發價格。 過去四十年來，台灣傳統農業面臨低生產率及缺乏國際競爭力的困境。台灣農業特性包含農業勞動者教育水準偏低和農業勞動人口逐年減少。 高麗菜是重要的民生必需品。蔬菜營銷問題一直是政府重要的課題。天然災害發生時，高麗菜零售價格會立即大幅上漲；相反的，當生產過剩時，市場價格時常會低於生產成本。這些現象促使本研究動機及引發預測高麗菜價格的興趣。 目前高雄果菜運銷公司高麗菜供應鏈面臨幾個問題。1)所有高麗菜供應鏈成員的菜價估計，全依賴個人的經驗及季節性因素。2)供應鏈成員缺乏一個科學的方法來分析和預測價格資訊。3)農委會2010年開始實施「大宗蔬菜種植登記暨供育苗量統計資訊系統」此資訊系統對高麗菜價格影響還是未知。 本研究整理高麗菜價格包含公開及私有資料共66周數據。私人數據是透過一位中盤商取得，從中盤商手寫筆記轉換成數位檔是一項大工程。我們目標是預測下周高雄果菜運銷公司高麗菜批發價格。本研究建構各種預測模型，其中包含公開資料及私有資料的各種組合，並評估各模型的預測準確性及誤差。我們發現，私有數據是非常有價值的，這對持續蒐集私人數據的高麗菜商是莫大的鼓舞。 關鍵字：時間序列預測、高麗菜價格、私有資料、高雄果菜運銷公司、台灣農業經濟、市場通路、價格傳遞、供應鏈、資料探勘。
並列摘要
This research aims to discover the value of private data on demand and supply, as collected by middleman in the cabbage supply chain in Taiwan, for the purpose of forecasting weekly cabbage prices. Low productivity and lack of global competitiveness are two predicament traditional agriculture facing during past 40 years. Taiwan’s agriculture is unique, and its characteristics include agricultural workers with low educational levels and a decreasing population of agriculture workers. Cabbage is an important crop in Taiwan. The cabbage marketing issue is important to the government. When a disaster occurs, vegetable retail price often suddenly upsurges, as has occurred in recent years. Conversely, when overproduction occurs, the market price is lower than production cost. These phenomena motivated this study and the need for price forecasting of cabbage. Current difficulties that cabbage supply chain members in Kaohsiung Agricultural Products Marketing Company face are: 1) the estimated price of vegetables used by members of the supply chain still relies on personal experience and seasonal factors. 2) Supply chain members lack a scientific method to analyze and forecast pricing information. 3) In 2010, the government launched a new policy “The Database of Bulk Vegetables Planting, Registration and Seedlings Supply”. The effectiveness of this system for cabbage pricing is still unknown. We obtained public and private data on cabbage prices for 66 weeks of data. The private data were obtained from a middleman and were converted from handwritten notes to digitized form. We focuses on forecasting weekly cabbage price, one a week ahead, in Kaohsiung, Taiwan. We fit models with and without the private data and evaluated one-week-ahead forecast accuracy. We find that private data is valuable. It means that private data improves forecasting performance. This is a great encouragement for the stakeholders in cabbage supply chain who continue to collect data. Keyword: Time Series Forecasting, Cabbage Price, Private Data, Kaohsiung Market., Taiwan’s Agriculture Economic, Marketing Channel, Price Distribution, Supply Chain, Data Mining.