論文中我們對海外基金提出了一種投資策略。我們先定義全球趨勢指標 (GTI)，來評定基金在價格變動上的未來趨勢。之後，以GTI為基礎，延伸出監控指標 (MI) 來測量基金市場是處於空頭市場或是多頭市場。MI決定買賣訊號。一個基金好壞是由短期與長期績效指標之權重組合來衡量。各種績效指標的權重是由基因演算法來做最佳化，其中基因演算法可以根據過去的績效 (獲利能力) 來動態調整基金好壞的優先權。當MI指標觸發了買進訊號，績效好的基金會比績效差的基金更有機會被挑選出來。從1999年1月到2008年12月 (總共10年)的實驗結果中，我們獲得高於10%年化報酬率，相對於其他已存在的方法，這是一個很顯著的改進。

In this thesis, we propose an investment strategy for the world mutual funds. We first define the global trend indicator (GTI) for evaluating the price change trend of the funds in the future. Then, based on GTI, we derive the monitoring indicator (MI) to measure whether the fund market is in the bull or bear state. MI decides the buying or selling signal. The goodness of a fund is determined by some weighted combination of short-term performance and long-term performance. The weights of various performances are decided by the genetic algorithm, which can dynamically adjusted with priorities of interested funds according to their past performances (profitability). When a buying signal is triggered by MI, the funds with high performance are more likely to be picked than those with low performance. In our experimental results from January 1999 to December 2008 (10 years in total), we achieve the annual profit higher than 10%, which is a significant improvement to other existing methods.

Chapter 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Chapter 2. Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1 The ROI and Average Annualized Return . . . . . . . . . . . . . . . 4
2.2 The 4433 Rule for Choosing Funds . . . . . . . . . . . . . . . . . . . 8
2.3 Data Collection and Preprocessing . . . . . . . . . . . . . . . . . . . 9
2.4 Technical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4.1 SMA (Simple Moving Average) . . . . . . . . . . . . . . . . . 10
2.4.2 EMA (Exponential Moving Average) . . . . . . . . . . . . . . 11
2.4.3 MACD (Moving Average Convergence/
Divergence) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.5 Genetic Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Chapter 3. Fund Investment with Genetic Algorithms . . . . . . . . 21
3.1 The Flow Chart of Investment . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Global Trend Indicator . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3 Evaluation of the Financial Market . . . . . . . . . . . . . . . . . . . 25
3.3.1 Monitoring Indicator . . . . . . . . . . . . . . . . . . . . . . . 26
3.3.2 The MACD Indicator . . . . . . . . . . . . . . . . . . . . . . . 35
3.4 Fund Selection with Genetic Algorithms . . . . . . . . . . . . . . . . 35
3.4.1 The Scoring Function . . . . . . . . . . . . . . . . . . . . . . . 35
Page
3.4.2 The Specification of the Genetic Algorithm . . . . . . . . . . . 36
3.4.3 The Procedure of Each Generation in GA . . . . . . . . . . . 37
Chapter 4. Experimental Results . . . . . . . . . . . . . . . . . . . . . . 41
4.1 Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.1.1 The Performance of the Motley Fool Investment . . . . . . . . 42
4.1.2 The Performance of the 4433 Rule . . . . . . . . . . . . . . . . 43
4.1.3 The Performance of the Buy-and-hold Strategy for Ten Years
with Maximal ROIcum . . . . . . . . . . . . . . . . . . . . . . 44
4.1.4 The Performance of the One-year Time Deposit . . . . . . . . 48
4.1.5 The Performance of the MSCI World Price Index . . . . . . . 48
4.1.6 The performance of the S&P 500 Composite Price Index . . . 49
4.2 Experimental Results with User-Defined Constant Weights . . . . . . 52
4.3 Experimental Results with Dynamic Weights Based on Genetic Al-
gorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4 Experimental Results with Constant Weights Based on Genetic Al-
gorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.5 Performance Comparison of Various Models . . . . . . . . . . . . . . 77
Chapter 5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.1 Conclusion for Our Experiments . . . . . . . . . . . . . . . . . . . . . 79
5.2 Further Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Appendixes
A. The Transaction Information of Our Best ROIcum . . . . . . . . . . 83

[1] R. Choudhry and K. Garg, “A hybrid machine learning system for stock market forecasting,” Proceedings of World Academy of Science, Engineering and Technology, Vol. 29, Copenhagen, Denmark, pp. 315–318, May 2008.
[2] FundDJ Co., Ltd., “FundDJ.” http://www.funddj.com/, 2000.
[3] J. H. Holland, Adaptation in Natural and Artificial Systems. University of Michigan Press, Michigan, 1975.
[4] W. Huang, Y. Nakamori, and S.-Y.Wang, “Forecasting stock market movement direction with support vector machine,” Computers and Operations Research, Vol. 32, No. 10.
[5] H. Ince and T. B. Trafalis, “Kernel principal component analysis and support vector machines for stock price prediction,” IEEE International Joint Conference on Neural Networks, Vol. 3, Budapest, Hungary, pp. 2053–2058, July 2004.
[6] H. Ince and T. B. Trafalis, “Short term forecasting with support vector machines and application to stock price prediction,” International Journal of General Systems, Vol. 37, No. 6, pp. 677–687, Dec. 2008.
[7] K.-I. Kainijo and T. Tanigawa, “Stock price pattern recognition: a recurrent neural network approach,” Proceedings of the international joint conference on neural networks, Vol. 1, Washington, USA, pp. 215–221, June 1990.
[8] K.-J. Kim, “Financial time series forecasting using support vector machines,”Neurocomputing, Vol. 55, pp. 307–319, Mar. 2003.
[9] K.-J. Kim and I. Han, “Genetic algorithms approach to feature discretization in artificial neural networks for the prediction of stock price index,” Expert Systems with Applications, Vol. 19, pp. 125–132, Aug. 2000.
[10] T. Kimoto and K. Asakawa, “Stock market prediction system with modular neural networks,” Proceedings of the International Joint Conference on Neural Networks, Vol. 1, Washington, USA, pp. 1–6, June 1990.
[11] P.-F. Pai. and C.-S. Lin, “A hybrid ARIMA and support vector machines model in stock price forecasting,” OMEGA: The International Journal of Management Science, Vol. 33, pp. 497–505, Dec. 2005.
[12] N. Powell, S. Y. Foo, and M. Weatherspoon, “Supervised and unsupervised methods for stock trend forecasting,” Proceedings of the 40th Southeastern Symposium on System Theory, New Orleans, USA, pp. 203–205, Mar. 2008.
[13] H. Someya and M. Yamamura, “Genetic algorithm with search area adaptation for the function optimization and its experimental analysis,” Proceedings of the 2001 Congress on Evolutionary Computation, Vol. 2, Seoul, South Korea, pp. 933–940, May 2001.
[14] StockCharts.com, “ChartSchool.” http://stockcharts.com/, 1999.
[15] T. Z. Tan, C. Quek, and G. S. Ng, “Brain-inspired genetic complementary learning for stock market prediction,” Proceedings of IEEE Congress of Evolutionary Computing, Vol. 3, Edinburgh, UK, pp. 2653–2660, Sep. 2005.
[16] F. E. Tay. and L. Cao, “Application of support vector machines in financial time series forecasting,” OMEGA: The International Journal of Management Science, Vol. 29, No. 4, pp. 309–317, 2001.
[17] F. E. Tay. and L. Cao, “Modified support vector machines in nancial time series forecasting,” Neurocomputing, Vol. 48, pp. 847–861, 2002.
[18] B. Taylor, “GlobalFinancialData.” https://www.globalfinancialdata.com, 1990.
[19] J. C. Wen, “The relationship between the investment behavior of institutional investors, interest rate, foreign exchange rate, the performance of US stock market and the return of Taiwan stock market,” Master’s Thesis, Department of Finance, Shih Hsin University, Taipei ,Taiwan, July 2007.