English  |  正體中文  |  简体中文  |  Items with full text/Total items : 26988/38789
Visitors : 2347018      Online Users : 32
RC Version 4.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Adv. Search
LoginUploadHelpAboutAdminister

Please use this identifier to cite or link to this item: http://ntour.ntou.edu.tw:8080/ir/handle/987654321/12969

Title: 台灣吳郭魚價格預測方法有效性之比較研究—類神經模糊理論之應用
Comparative Study of Price Forecasting Methods for Taiwan Tilapia—An Application of Neuro-Fuzzy
Authors: Wei-Chun Fu
傅偉鈞
Contributors: NTOU:Institute of Applied Economics
國立臺灣海洋大學:應用經濟研究所
Keywords: 迴歸模型;GARCH;類神經模糊;RMSE;KD指標
Regression model;GARCH;Neuro Fuzzy;RMSE;KD index
Date: 2004
Issue Date: 2011-06-30T07:09:24Z
Abstract: 本研究主要探討類神經模糊理論在吳郭魚批發價格預測上的應用。以往眾多學者對於吳郭魚價格預測的研究,大多以線性預測模式(包括計量經濟模型、時間序列模型及組合預測)方法進行分析。線性模型因受限於基本的假設條件,且須考量許多影響的因素,在進行魚價預測上花費了許多的成本,使得預測變得極為複雜且不經濟。再者,吳郭魚價格的變動,受可預料或無法預料的因素所影響,加上吳郭魚供需力量無法平衡的關係,使得魚價波動更不穩定,造成業者損失慘重。因此,本研究建構人工智慧類神經模糊理論的預測模型,應用在魚價漲跌幅度的預測,是有其必要性及迫切性。 本研究使用兩種建構預測模型的方式,一是利用價格及數量關係建構迴歸、GARCH(1,1)、GARCH-M(1,1)與價量三次非線性模型;另ㄧ是KD指標方式建構迴歸、GARCH(1,1)、GARCH-M(1,1)與類神經模糊模型,並將資料分成訓練集與驗證集資料,利用訓練集資料建構所需要的模型,最後將驗證集資料帶入模型以求證模型的優劣,分別以Theil 不等係數U值、均方根誤差(RMSE)與預測方向正確率三種指標作為衡量預測模型優劣的基準。模擬結果顯示,價量所建構的預測模型,以非線性價量模型預測能力較佳;KD指標方式所建立的預測模型,以非線性類神經模糊模式預測能力較佳,由此可推論非線性之預測模型較適合預測吳郭魚價格之漲跌幅,其中以KD指標建構預測模型之預測能力較優於價量關係所建構之模型。
The purpose of this research is to explore that Neuro Fuzzy to price forecasting for Taiwan Tilapia. Most of the scholarly researches are regarding price forecasting that are used to linear models that it is limited to conditions of basic hypothesize and many factors of effecting price, and it becomes more complex and not economical. Furthermore, the Tilapia price would take effect by expected or not expected factors and supply and demand unable equilibrium, and it makes Tilapia price fluctuation instable. Thus this research to construct Artificial Neuro Fuzzy model to application of price forecasting for Tilapia, and it would be desirable. This research use two constructing ways of forecasting models, one way of using the relation of price and quantity and construct regression model, GARCH(1,1) model, GARCH-M(1,1) model, and nonlinear forecasting model. The other way of using KD index and construct regression model, GARCH(1,1) model, GARCH-M(1,1) model, and nonlinear Artificial Neuro Fuzzy model and use three indexes that Theil’s inequality coefficients U value, Root Mean Square Error(RMSE), and correct rate to judge those models are better or worse. The following results are found: This research’s nonlinear forecasting model in two constructing ways of forecasting models are both better than linear models, especially Neuro Fuzzy is better than others.
URI: http://ethesys.lib.ntou.edu.tw/cdrfb3/record/#G0M92350003
http://ntour.ntou.edu.tw/ir/handle/987654321/12969
Appears in Collections:[應用經濟研究所] 博碩士論文

Files in This Item:

File Description SizeFormat
index.html0KbHTML203View/Open


All items in NTOUR are protected by copyright, with all rights reserved.

 


著作權政策宣告: 本網站之內容為國立臺灣海洋大學所收錄之機構典藏,無償提供學術研究與公眾教育等公益性使用,請合理使用本網站之內容,以尊重著作權人之權益。
網站維護: 海大圖資處 圖書系統組
DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback