透過您的圖書館登入
IP:3.19.31.73
  • 學位論文

以資料探勘法驗證股價的可預測性

Apply Data Mining on the Predictability of Stock Price

指導教授 : 呂麒麟 吳樹欉
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


過去預測股價的指標大致上可分為基本分析指標與技術分析指標兩種,近年來發現股權結構也有顯著的價格影響能力,卻少有研究將股權結構與價格相關指標結合對台灣股價預測做探討。本研究同時納入基本分析指標、技術分析指標與股權結構指標對股價做預測,研判是否能更準確的預測股價變動前的準則或徵兆,同時並使用近年來廣泛運用在預測上面的資料探勘技術,驗證是否能藉由新技術來提高股票變動前的預測能力並歸納其規則性,以作為投資人在選購企業股票與時點上之評估依據。實證結果顯示,決策樹與類神經網路在不同時點下,決策樹在大漲前五天的準確率優於前ㄧ天及當天,類神經網路則相反,因此同時使用決策樹與類神經網路做預測,最有利於投資人擬訂較正確的選股策略。

並列摘要


The literature of predict stock price can be broadly divided into the fundamental analysis and technical analysis. In recent years ownership structure found has significant impact on the stock price, but only few study discuss ownership structure and stock predict indicators together to forecast the price change in Taiwan. This research incorporate the indicators of fundamental analysis, technical analysis and ownership structure to predict the stock price change. I employ draw Data Mining Technology, which were widely used in prediction, to confirm that we can use it to enhance the ability of forecast before price change. This study shows that Decision Tree and Artificial Neural Network in Data Mining have different effect on time. Decision Tree has better accuracy at the day before the price rise, Artificial Neural Network vice versa. It’s more advantage for investors to make the investment decisions with both Decision Tree and Artificial Neural Network at the same time.

參考文獻


ㄧ、中文部分
1. 李明黎(2006),第一次股票低買高賣就上手(技術面+籌碼篇),易博士文化出版:城邦文化發行。
2. 陳共,周升業,吳曉求(2001),證?投資分析,五南圖書出版公司。
3. 林傑斌,劉明德,陳湘(2002),資料採掘與OLAP理論與實務,文魁出版與發行。
4. 孫惠民(2007),資料採掘理論與實務規劃手冊,文魁出版與發行。

延伸閱讀