- 學位論文
應用灰色理論預測汽車產業銷售-以南部某汽車經銷商為例
Grey predication Theory on Automobile Sales Forecasting - A Case Study of the Dealer Company in Southern Taiwan
摘要
處於雲端科技發達時代中，從網路取得各項所需的參考資料並非難事。但如何將資料（Data）轉換歸納出有效的資訊（Information），或則將資料中的雜訊篩濾出來，使決策者準確作出關鍵的決策，仍是現今學術界重要探討之議題。故預測科學各種不同的預測模型如雨後春筍產生，而模型各具特色及其適切的資料預測型態，例如，近代的灰色預測模型，因只需四筆資料便可作預測，其特性已被廣泛運用於各種研究領域之中，就以GM（1,1）的應用較為廣泛。 基於台灣汽車市場需求受限於人口規模，大幅成長空間不高，但整體汽車產業涵蓋電子、塑膠、鋼鐵、橡膠、石油等產業供應鏈，屬資本及技術密集產業，對於國家整體經濟貢獻度是不容小覷，汽車產業銷售概況可作為國家經濟成長參考指標之一，如能精準預測未來銷售趨勢，對於該產業營運決策勢必能給予相當助益。 本研究主要探討汽車銷售預測，以南部某汽車經銷商之銷售預測作為研究標的，運用灰色關聯模型分析總體經濟指標其中經濟成長率(Economic Growth Rate, EGR）、國民所得(National Income）、消費者物價指數(Consumer Price Index, CPI）、美元匯率(USD exchange rate）、國民生產毛額(Gross National Product, GNP）、貨幣總計數M2(Monetary Aggregates）(M2）、失業率(Unemployment Rate）、儲蓄率(Savings Rate）、投資率(Investment Rate）及等，探討影響整體汽車銷售量關聯排序，再以個案汽車經銷商2012-2015年季銷售量作為研究對象，以灰色理論模型GM（1,1）、移動平均法、平滑指數及線性迴歸預測模型四種預測方式做分析比較，進行汽車銷售量之預測。 經研究結果顯以經濟成長率、國民所得、消費者物價指數、美元匯率等項與汽車銷售最具關聯性的影響因素，並以GM（1,1）、移動平均法、平滑指數及線性迴歸預測模型四種預測模型比較之後，以GM（1,1）的預測最準確，未來仍呈現平穩成長趨勢，並依結果提出論述及相關建議，供後續相關研究之探討參據。
並列摘要
In the era of cloud computing technology developing, it is not difficult to retrieve data from the internet. The critical issue is, however, how to transfer “Data” into useful “Information”, or to remove the noises, in order to provide leadership for decision-making. Therefore, the various kinds of forecasting science models were developed, and each model has its feature and feasible data format. For example, the Grey Prediction Theory has been vastly applied in many research fields, because it only needs 4 source data to perform prediction, especially the GM (1,1). Taiwan automobile market is constrained by the population scale and hard to have a huge development. However, the automobile industry includes the supply chains of electronics, plastics, steels, petroleum, and etc. It belongs to the capital and technology intensive industry, and contributes a great shares of national gross economy. This research focused on the prediction of automobile sales, took a southern automobile dealer company as an example, and utilized Grey Relational Model to analyze Macroeconomic Indices, such as Economic Growth Rate, National Income, Consumer Price Index, USD Exchange Rate, Gross National Product, Monetary Aggregates (M2), Unemployment Rate, Savings Rate, Investment Rate, and to discuss the influences on automobile sales. Then utilized and compared the Grey Theory Models GM (1,1), Moving Average, Exponential Smoothing, and Linear Regression to perform market sales forecasting, based on the automobile dealer company's sales values from 2012 to 2015. The research results indicated that Economic Growth Rate, National Income, Consumer Price Index, and Dollar Exchange Rate were the most influential factors related to the automobile sales. In addition, compared the four prediction models, such as GM (1,1), Moving Average, Exponential Smoothing, and Linear Regression, GM (1,1) produced the most precise forecasting, and presented a steady and smooth development in the future. Based on the research results, we provided some recommendations for the reference of following discussion.