以資料採礦技術探討影響個股績效之關鍵因素 -以台灣上市個股為例-

INVESTIGATING THE CRITICAL FACTORS OF STOCK INVESTMENT PERFORMANCE OF TAIWAN STOCK MARKET THROUGH DATA MINING

甘堯昌(Yao-Chang Kan)

指導教授 ： 陳煇煌

大同大學/經營設計學院/資訊經營學系所/碩士(2007年)

摘要

在低利率時代裡，如何保有資金不被通貨膨脹折價？因此如何投資就變成大家的課題。為了尋求更高的報酬率，投資台股幾乎成為全民的運動。股票投資又是眾人又愛又恨，愛之可觀的報酬，恨之無法掌控的不確定性。逃避不如面對，故如何了解台灣股市個股漲跌的關鍵因素就變得極為重要了。
要在股市裡想要得到良好的績效，最快的方法就是學習頂尖專家的法則。華倫.巴菲特，世界公認藉由投資成為世界排名第一的富翁(2002,富比士)。故本研究整理華倫.巴菲特投資法則，承襲其精神，將探討評估企業價值與看待市場價格這二構面。
而隨著時代的進步，推陳出新的方法與工具，亦產生了許多新的構面來看待「股市」。資料採礦(Data Mining)技術是近年來被廣泛應用於各領域來協助分析大量資料的技術，主要的訴求希望透過資料採礦技術，從大量的歷史資料中挖掘出其中富有價值的知識或訊息。本研究將以資料採礦的工具，結合不同的思維來探討台灣上市個股績效的影響因素。
透過本研究改善其投資法則後，確實改善了原先的累增報酬率及年複利率。且可以確定投資台灣股市個股亦可以達到華倫.巴菲特的年複利率22.37%。本研究模擬5年期的投資行為後，最高可得累增報酬率及年複利率為1067.74%、63.48%。

關鍵字

類神經網路 ； 價值投資法則 ； 資料採礦 ； 台灣股市 ； 決策樹

並列摘要

How to keep your asset without being devaluated by inflation at the time of low interest rate? Investment is a very important issue to everyone. To obtain high rewards, it is popular that people make an investment in the Taiwan Stock Market. However, people love and hate it for its remarkable returns and its risks. It is crucial to understand the key factors that influence the fluctuation of stock value.
To obtain the best yield from the stock market, the fastest way is to learn the expert’s tactics. Ranked by Forbes in 2002, Warren Buffertt is the richest man by investing in the world. This research will thus follow his investment principles and probe the intrinsic values of enterprises and market values of stock shares.
New analytical methods and tools on stock market have been developing and make us look on the market in different aspects. ‘Data mining’ is the method that has been applied widely in different fields to analyze a large amount of data in recent years. Its main purpose is to get the useful information and important message about stocks from their historical records. This research combines different assumptions with Data Mining skills to study the critical factors that affect the stock returns.
Adjusting the investment principles by data mining technique improves the original accumulative rewarding rate and yearly compound interest rate. It is also assured that the stock returns in Taiwan market can reach up to 22.37%, the same annual compound interest rate as Warren Buffertt has obtained. We calculate the return of the investment in 5 years can be up to 1067.74% of accumulative reward and 63.48% of compound interest rate.

並列關鍵字

Data Mining、Regression Trees、Artificial Neural

參考文獻

7.呂宗耀、張献祥(民94)，「投資者的修練」，時報文化出版企業股份有限公司。

22.鄭弘儀(民92)，「鄭弘儀教你投資致富，高富國際文化股份有限公司出版。

18.葉勁之(民91)，「財務規劃顧問的第一本書」，城邦文化事業股份有限公司出版。

8.李惠妍(民90)，「類神經網路與迴歸模式在台股指數期貨預測之研究」，國立成功大學管理學院碩士論文。

13.俞海琴、葉宜生(民90)，「價值投資法於台灣股市適用性之研究-電子產業與非電子產業之配對比較」，中原大學國貿學系碩士論文。