使用基因演算法以預測台股加權指數

Using Genetic Algorithms to Forecast Taiwan Stock Market Trend

王義昌(YI-CHANG WANG)

指導教授 ： 林志浩

中原大學/商學院/資訊管理學系/碩士(2008年)

https://doi.org/10.6840/cycu200900593

摘要

在台灣，許多人以股票市場作為投資理財的管道之一。雖然股市具高報酬的特性，但也同時伴隨著高度的風險性，因此想要靠投資股市獲利，絕不是一件容易的事。或許大多數人都知道要研究經濟基本面與股市技術面，但是面對龐雜而繁瑣的資訊，卻也令許多人望而却步，因此，一個適當的輔助工具是絕對必要的。
技術指標是大家常聽到也是常用到的工具之一，許多股票投資人及分析師都會使用技術指標來分析股市，雖然用的人很多，但是如何用得好也用得巧卻不是很容易。技術指標多達數十種，有共線性高者，也有買賣訊號互為矛盾者，當數種指標結合在一起分析後，常會讓人感到混淆而不知所措，縱有再多指標也無法發揮預測的效果。有鑑於此，本研究之目的即在發展一個預測模型，以常見的18種技術指標作為輸入因子，利用基因演算法之強大搜尋能力，試圖在看似雜亂無章的一堆數據中，找尋出最佳之技術指標組合，再套入預測模型中以預測次日收盤價。
觀諸過去之許多研究文獻，有些採用單一方法以預測股市，也有一些結合數種方法來預測股市，但方法往往過於複雜，不易理解與應用。相較於其他模型，本模型最大的優點是簡單、易於理解，且從1999年至2003年之實驗比較中，證實本模型同時具有相當不錯的預測準確度。同時，本研究也針對演化世代對預測結果的影響進行實驗與分析，發現演化世代不宜太大，以免降低預測的能力，期望本研究對後續之研究者能有些許幫助。
關鍵字：基因演算法，技術指標，技術分析，股市預測

關鍵字

基因演算法 ； 技術指標 ； 技術分析 ； 股市預測

並列摘要

The stock market is one of the main ways in which Taiwanese people choose to invest their money. However, although there is the potential to earn high returns, making a profit from the stock market is no easy matter. While most investors recognize that they must study both fundamental and technical analysis, many people can be left feeling helpless in the face of the sheer volume and detail of the information available. Therefore, it is vital for investors to find a tool that is appropriate for their needs.
Technical indicator may be a tool widely used by stock market investors, but using it well is a different matter entirely. This study provides a prediction model, which takes 18 commonly used types of technical indicators as input factors. It uses Genetic Algorithms to obtain the optimum combination of technical indicators, which are then used in the prediction model to forecast the next day’s closing price.
In past studies, sometimes a single method would be used, and other times a combination of several approaches would be employed for stock prediction, but too often these would all be too complex and difficult to use. The biggest advantage of this model is that it is simple and easy to understand, and, furthermore, actual test results from 1999 to 2003 have confirmed that this model has an excellent accuracy rate. As a result, it is hoped that this study can be of great benefit to researchers.

並列關鍵字

Technical Indicators ； Genetic Algorithms ； Stock Prediction ； Technical Analysis

參考文獻

[1]. K. Kim, I. Han, (2000), “Genetic algorithms approach to feature discretization in artificial neural networks for the prediction of stock price index,” Expert Systems with Applications, Vol. 19, Issue 2, pp. 125-132.

[2]. T. Kimoto, K. Asakawa, (1990), “Stock Market Prediction System with Modular Neural Networks,” IEEE International Joint Conference on Neural Networks, Vol. 1, pp. 1-6.

[3]. S. C. Chi, W. L. Peng, P. T. Wu, and M. W. Yu, (2003), “The study on the relationship among technical indicators and the development of stock index prediction system,” IEEE International Conference of the North American on Fuzzy Information Processing Society, pp. 291-296.

[4]. M. Dash, H. Liu, (1997), “Feature Selection for Classification,” Intelligent Data Analysis – An International Journal, pp. 131-156.

[5]. D. Enke, S. Thawornwong, (2005), “The use of data mining and neural networks for forecasting stock market returns,” Expert Systems with Application – An International Journal, Vol. 29, Issue 4, pp. 927-940.