應用馬可夫決策過程與遺傳演算法於台灣股市投資策略制訂

Using Markov Decision Process and Genetic Algorithms for Formulating Taiwan Stock Trading Strategies

李明昇(Ming-Sheng Lee)

指導教授 ： 張應華

淡江大學/管理學院/資訊管理學系碩士班/碩士(2012年)

https://doi.org/10.6846/TKU.2012.01190

摘要

隨著低利率時代來臨，投資者為了追求較高的報酬率，開始把資金投入股票投資市場，然而股市行情變換迅速，真正獲利的投資者不多，只有在適當時機點進場交易的投資者才能從中獲利。ㄧ般投資者大多利用技術指標做為進場時機的依據，然而使用技術指標會有ㄧ些問題，例如技術指標的選擇、互相矛盾或類似等問題，導致ㄧ般投資人很難利用這些資訊來輔助股市投資決策。
本研究結合馬可夫決策過程與遺傳演算法，提出新的分析架構，建立一個股市投資策略制訂的決策支援系統。本研究利用馬可夫決策過程具有的預測特性和近期資料即時分析能力，利用馬可夫決策過程在短期資料優秀的分析能力，解析過去歷史資料，達到擇時效果，再結合遺傳演算法的特殊編碼方式與極佳搜尋能力，以字串編碼表達不同的投資策略，以搜尋能力來求解出最佳投資策略，達到選股和資金配置效果。在資金與手中持股不足時，可透過此模型具有的融資融券方式來完成交易。經實驗證實本模型可以得到較高的報酬。

關鍵字

馬可夫決策過程 ； 遺傳演算法 ； 股市投資 ； 擇時 ； 最佳化

並列摘要

With the low interest rate coming, investors start to buy stocks to get more rewards. However, the stock market varied rapidly, seldom investors can get excess returns when trade in the proper time. Most investors use technical indicators as a tool for market timing. However using technical indicators has some problems, such as the choice of technical indicators, conflicting or similar and other prolems. So most investors are difficult to use those informations to determine stock market investment decisions.
This research combines Markov decision process and genetic algorithms to propose a new analytical framework and to develop the decision support system for making the stock trading strategies. This paper uses the prediction characteristics and real-time analysis capabilities of the Markov decision process to do timing decision. Doing the stock selection and fund allocation by using the string encoding to express different investment strategies and the search capabilities to solve the best investment strategy. Besides, when investors have no sufficient money and stocks, the architecture of this research can complete the transaction by credit transactions. By the experiments, it can confirm that the model of this research can get higher reward.

並列關鍵字

Markov decision process ； genetic algorithms ； the stock investment ； timing ； optimization

參考文獻

[4] 姚凱齡，〈整合限制滿足概念於遺傳演算法制定股票投資策略之研究〉，大同大學資訊經營研究所碩士論文，2008。

[12] Allen, F., Karjalainen, R., “Using genetic algorithms to and technical trading rules”, Journal of Financial Economics, vol. 51, no. 2, 1999, pp. 245-271.

[13] Bao, Y., “Stock market prediction model based on genetic algorithm and support vector regression”, Proceeding of the International Conference on Energy and Environmental Science, Singapore, 2011, pp. 4025-4029.

[15] Bermúdez, J. D., Segura, J. D., Vercher, E., “A multi-objective genetic algorithm for cardinality constrained fuzzy portfolio selection”, Fuzzy Sets and Systems, vol. 188, no. 2, 2012, pp. 16-26.

[16] Chang, T. J., Yang, S. C., and Chang, K. J., “Portfolio optimization problems in different risk measures using genetic algorithm”, Expert Systems with Applications, vol. 36, no. 7, 2009, pp. 10529-10537.