運用基因演算法建立不同時期之最佳選股模式

The use of genetic algorithms to create different times of the best stock selection model

康譽騰(YU-TENG KANG)

指導教授 ： 葉佳炫

元智大學/資訊學院/資訊管理學系/碩士(2012年)

摘要

隨著資訊科發展，投資方式也跟著變多，在眾多的投資方式中，股票仍是大多數人所選擇的投資方式，研究目的是在股市不同時期找出最佳的投資策略。研究先以基本面選擇標的物，對象為上市的半導體產業，找出該產業較有發展的公司，因基本面能評估該公司體質的好壞。因此研究中從過往的論文中，找出對股價有正相關的因子，做為評判是否選取的標準，而所選的基本面分別為毛利率、股價淨值比、現金流量和本益比，利用四基本面找出最佳的投資標的。在交易策略中，利用技術指標作為進出場的依據，主要是讓投資人了解技術指標的有效性。研究中將從過往的論文中，對使用過的技術指標做統計，找出最常用的四項技術指標，並結合基因演算法做交易策略的訓練、測試，將結果和買入持有的投資方式做比較。實驗中設定不同的訓練期和測試期，並分析基因演算法產生的交易策略適合何種交易方式，來幫助投資者建立交易策略，在本研究發現，對於股市進出場投資判斷，建議投資者採用訓練期一個月和測試期一個月會有較佳的投資報酬表現及較低的投資損失風險，而在面對空頭市場時，基因演算法產生的交易策略能較買入持有得到更好的績效。

關鍵字

基因演算法 ； 基本面 ； 技術分析 ； 投資組合 ； 選股 ； 擇時

並列摘要

With the development of IT, the investing ways are getting more and more. Among many types of investment, the stock is still the one that most people choose. The purpose of the study is to find out the best investing strategy in different periods of stock market. First of all, by the fundamentals, we choose semiconductor industry as the target and find out the company which is more potential in development. The reason is that the fundamentals can be used to estimate the quality of companies.
Therefore, from the literature, we find out positive factors which affect the stock price as the standard to select the candidate companies. The fundamentals include the ratio of margin, the price to earnings patio, the price-book ratio, and the cash flow. The four fundamentals are used to discover the best targets. In the trading strategy, we utilize the technical indicator as the basis of buying and selling. The purpose is mainly to allow investors to understand the effectiveness of technical indicators. Besides, from the previous literature, we select four technical indicators which people used most frequently, and then combine them with the genetic algorithms to train and test the trading strategies. We compare the results with the buy-and-hold strategy.
During the experiments, we set different training periods and testing periods. Then, we analyze which trading way the strategies generated by means of the genetic algorithm fit for. In this way, we could help investors search better trading strategies. Moreover, we recommend investors to select the one-month training and one-month testing period. It would profit more and lower the risk of investment losses. Consequently, when facing the bear market, it would be better to choose the strategies generated from the genetic algorithms rather than the buy-and-hold strategy.

並列關鍵字

Stock Selection ； Technical Timing ； Genetic Algorithms ； Technical Analysis ； Portfolio

參考文獻

[22]謝宏鑫，「考量投資標的的基本面、技術面分析並運用遺傳基因演算法決定最佳投資組合」，臺灣台北，台北大學企業管理學系碩士論文，2009。

[1]Shian-Chang Huang and Tung-Kuang Wu (2008), "Integrating GA-based Time-Scale Feature Extractions with SVMs for Stock Index Forecasting", Expert Systems with Applications (SCI, EI), 35 (4), pp. 2080-2088, November 2008.

[2]Chan,Louis K.C.,Yasushi Hamao and Josef Lakonishok, "Fundamentals and stock returns in Japan, " Journal of Finance 46,1739-1789, 1991.

[3]Fama, E.F., French, K.R., "The cross section of expected stock returns," Journal of Finance 47, 427-466,1992.

[4]LinCaoWangZhang, "The Applications of Genetic Algorithms in Stock Market Data Mining Optimisation", 2000.