標題: 艾略特波浪理論之股市趨勢分析-結合自組織映射圖神經網路與倒傳遞類神經網路
Analyzing Taiwan Stock Trend Based on Elliott Wave Theory—Combining Self-Organizing Map and Back- Propagation Neural Network
作者: 江松茂
Chiang, Sung-Mao
陳安斌
Chen, An-Pin
管理學院資訊管理學程
關鍵字: 艾略特波浪理論;技術分析;自組織映射圖神經網路;倒傳遞類神經網;Elliott Wave Theory;Technical analysis;Self-organizing map neural network;Back-propagation neural network.
公開日期: 2010
摘要: 艾略特波浪理論之股市趨勢分析-結合自組織映射圖神經網路與倒傳遞類神經網路 學生:江松茂 指導教授:陳安斌 博士 管理學院(資訊管理學程)碩士班 摘要 台灣股市為淺碟型市場,近年國際金融局勢不穩,台灣股市也隨著齊漲齊跌,投資人一旦看錯方向就會遭受到巨大的損失,如何判斷所處時空環境中屬於漲勢、跌勢是一個重要的議題。 技術分析因資料取得及資料可定量,且相較於基本面分析,資料污染導致錯誤分析的機會較低,為大部份投資人常使用之投資分析工具。而艾略特波浪理論為近來廣泛研究之技術分析之一,洪若信(2009) 透過倒傳遞類神經網路進形波浪理論之參數學習、訓練,證明了股市中物理力道的存在。陳儒琦(2010) 篩選符合俱艾略特波浪理論特徵規則的N型波浪,可有效掌握股市大漲大跌走勢。本研嘗試以自組織映射圖神經網路之「物以類聚」特性,將波浪理論中參數依不同盤勢特徵加以分群,結合倒傳遞類神經網路學習、訓練並預測其漲跌幅,分析不同盤勢中的物理現象。 研究結果顯示,自組織映射圖神經網路可對波浪理論中N型波浪進行有效分群,且大跌之群集物理力量最為顯注。在考慮買入點為轉折點發現第三日收盤價情形下,大漲之群集買進成本相對轉折點最高。顯示投資人在大漲格局追價時應更加小心。 關鍵字:艾略特波浪理論、技術分析、自組織映射圖神經網路、倒傳遞類神經網路
Analyzing Taiwan Stock Trend Based on Elliott Wave Theory—Combining Self-Organizing Map and Back Propagation Neural Network Student:Sung-Mao Chiang Advisor: Dr. An-Pin Chen Institute of Information Management National Chiao Tung University Abstract Institute of Information Management National Chiao Tung University The stock market in Taiwan is a shallow-plate market which is affected by international stock market decline. In recent years, global economic is unstable. Investors will suffer great loss once they do the wrong direction. How to determine which space-time environment are upward, downward is an important issue. Technical analysis for quantitative data and information obtained and compared to fundamental analysis, data analysis of pollution leading to a lower chance of error, for the majority of investors often use investment analysis tool . The Elliott Wave Theory for the recent extensive study of technical analysis. Hong’s study (2009) through the back-propagation neural network parameters into the shape of the wave theory of learning, training, and that the stock market in the presence of physical strength. Chen (2010) filtering characteristics meet all the rules of Elliott Wave theory of N-type waves, which can effectively control the stock market rose fell trend. This study attempts to self-organizing map neural networks, "Like attracts like" feature, the wave theory of parameters to be characterized by different grouping plate potential, combined with back-propagation neural network learning, training and predict its decline in inflation, analysis of different plate potential the physical phenomena. The results show that self-organizing map neural network may be the wave theory of N-type waves for effective clustering and collapse of the power of the most obvious cluster physical note. Turning point in considering the purchase of the third day's closing price that case, the cluster rose a turning point in buying costs are relatively high. This study shows the pattern of investment in chase prices rose to be more careful. Keywords: Elliott Wave Theory、Technical analysis、Self-organizing map neural network、Back-propagation neural network.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT079864515
http://hdl.handle.net/11536/48625
Appears in Collections:Thesis