臺灣博碩士論文加值系統

在股票市場中，技術分析是常用來預測股價波動的方法之ㄧ。然而因為技術指標的複雜性及其多樣性，導致一般投資大眾不易使用技術指標來判斷股票的漲跌。此外，股市分析專家及基金經理人常會以自身主觀的判斷來運用技術指標，而對股市訊號做出錯誤的判斷。因此，此研究提出一多技術指標混合模式於投資決策上提供簡單且有效之規則。其模式包含下列三個階段：(1) 使用累積機率分配法及最小亂度法根據觀察值(技術指標值) 之資料分佈特性來對技術指標加以離散化，(2) 使用粗集理論來產生合理之規則， (3) 使用基因演算法來增進預測正確率及股票獲利率。在驗證本文所提模式之有效性上，採用台灣積體電路製造公司及台灣股市加權指數之股價資料，與粗集理論、基因演算法及「買進並持有」投資策略等模式來進行正確率及股票獲利之比較。實驗結果顯示，本文所提模式於正確率及股票獲利方面之績效優於其他模式。

In stock market, technical analysis is a major method for forecasting price fluctuation. However, it is hard for common investors to use this technique because there are too many different technical indicators to consider and too complex to understand. Besides, stock market analyst and fund managers forecast price fluctuation depend on their subjective judgments for technical indicators, which might lead to error judgment on market signals. Therefore, this study proposes a hybrid multi-technical indicators model, which combined three phases to provide easy and efficient rules for investment decisions: (1) consider the distribution characteristic of observations (the values of technical indicators) to partition each technical indicators by using cumulative probability distribution approach and minimize entropy principle approach, (2) produce linguistic rules by employing rough sets theory, and (3) utilize genetic algorithm to refine the generated rules, in order to improve forecasting accuracy rate and stock returns.
The effectiveness of the proposed model is verified by using a seven-year period of TSMC (Taiwan Semiconductor Manufacturing Company) stock price data and a six-year period of TAIEX (Taiwan Stock Exchange Capitalization Weighted Stock Index) stock index data, and comparing forecasting accuracy rate and stock returns with rough sets theory, genetic algorithm, and “Buy-and-Hold” approach. The experimental results show that the proposed model surpasses the other comparison models in accuracy rate and stock returns.

摘要 I
Abstract II
致謝 III
Contents IV
List of Tables V
List of Figures VI
1. Introduction 1
1.1 Research Background 1
1.2 Research Motivations and Objectives 2
1.3 Research field and Limitations 3
1.4 Organization of This Thesis 4
2. Related Works 5
2.1 Technical Analysis 5
2.2 Cumulative Probability Distribution Approach 7
2.3 Minimize Entropy Principle Approach 9
2.4 Rough Sets Theory 12
2.5 Genetic Algorithms 14
3. Methodology 18
3.1 Research Framework 18
3.2 Proposed Algorithm 21
4. Evaluations and Comparisons 26
4.1 Taiwan Semiconductor Manufacturing Company 28
4.2 Taiwan Stock Exchange Capitalization Weighted Stock Index 44
4.3 Findings and discussions 45
5. Conclusion 45
Reference 45
Appendix A. The economic meaning of Technical Indicators 45
Appendix B. The formula of Technical Indicators: 45
Appendix C. Correlations of Technical Indicators 45

[1]Acklam, P. J., 2004, An algorithm for computing the inverse normal cumulative distribution function.
[2]Allen, F. and Karalainen, R., 1999, Using Genetic Algorithms to Find Technical Trading Rules, Journal of Financial Economics 51, 245-271.
[3]Anna, B., 2007, Should normal distribution be normal? The Student’s T alternative, Computer Information Systems and Industrial Management Applications, 3-8.
[4]Azo, M.E., 1994, Neural Network Time Series Forecasting of Financial Markets, Wiley, New York.
[5]Blickle, T., and Thiele, L., 1995, A mathematical analysis of tournament selection, in Eshelman, L.J., Proceedings of the 6th International Conference on Genetic Algorithms, 506-511.
[6]Chen, A. P., Chang, Y. H., 2005, Using extended classifier system to forecast S&P futures based on contrary sentiment indicators, Evolutionary Computation, 2084 – 2090.
[7]Chen, A. P., Chen, Y. C., and Tseng, W. C., 2005, Applying Extending Classifier System to Develop an Option-Operation Suggestion Model of Intraday Trading-An Example of Taiwan Index Option, Lecture Notes in AI, 27-33
[8]Cheng, C. H., Chang, J. R., Yeh, C. A., 2006, Entropy-based and trapezoid fuzzification-based fuzzy time series approaches for forecasting IT project cost, Technological Forecasting & Social Change 73, 524–542.
[9]Cheng, C. H., Chen, J. S., 2008, Extracting classification rule of software diagnosis using modified MEPA, Expert Systems With Applications 34, 411-418

[10]Cheng, C. H., Yeh, C. A., 2004, New Fuzzy time series approaches for forecasting in expenditure of information project.
[11]Chi, S. C., Peng, W. L., Wu, P. T., and Yu, M. W., 2003, The study on the relationship among technical indicators and the development of stock index prediction system, Fuzzy Information Processing Society, 291 - 296
[12]Christensen, R., 1980, Entropy minimax sourcebook, Vol. 1-4, and Fundamentals of inductive reasoning, Entropy Ltd., Lincoln, MA.
[13]Clarence, N. and Tan, W., 1999, A Hybrid Financial Trading System Incorporating Chaos Theory, Statistical and Artificial Intelligence/Soft Computing Methods, School of Information Technology, Bond University, Queensland Finance Conference.
[14]Eshelman, L.J., Caruana, R.A., and Schaffer, J.D., 1989, Biases in the crossover landscape, in Proceedings of the 11th international joint Conference on Artificial Intelligence, Morgan Kaufmann, San Mateo, 10-19.
[15]Faerber, E., 1995, All about stocks: from the inside out, Chicago :Probus.
[16]Francis E.H. Tay *, and Lixiang S., 2002, Economic and financial prediction using rough sets model, European Journal of Operational Research 141, 641-659.
[17]Goldberg, D.E., 1989, Genetic Algorithms in Search, Optimization and Machine Learning, Reading, MA: Addison Wesley.
[18]Goldberg, D.E. and Deb, K., 1991, A comparative analysis of selection schemes used in genetic algorithm, Rawlins, G., (Ed.), Foundation of Genetic algorithms, Morgan Kaufmann, 69-93.
[19]Wang, Y. F., 2003, Mining stock price using fuzzy rough set system, Expert Systems with Applications 24, 13-23.

[20]Holland, J.H., 1975, Adaptation in Nature and Artificial Systems, University of Michigan Press.
[21]Kim, M. J., Min, S.H., and Han, I., 2006, An evolutionary approach to the combination of multiple classifiers to predict a stock price index, Expert Systems with Applications 31, 241-247.
[22]Li, J. B., Yu, Y. T., and Chen, A. P., 2006, Integration of Group Decisions and XCS in Intelligent Financial Decision Support System — an Example of Taiwan Index, Evolutionary Computation, IEEE Congress, 2389 – 2396.
[23]Pawlak, Z., 1982, Rough sets, International Journal of Computational Information Science, 341-356.
[24]Pawlak, Z. and Skoworn A., 2007, Rudiments of rough sets, Information Science, 177.
[25]Pring, M.J., 1991, Technical Analysis, New York.
[26]Reilly, F. K., 1989, Investment Analysis and Porfolio Management, Chicago, Dryden Press.
[27]Richard, J. B. and Julie, R. D., 1999, Technical market indicators :analysis & performance, New York :John Wiley.
[28]Ross, T.J., 2000, Fuzzy logic with engineering applications, McGraw-Hill, USA.
[29]Schulenburg, S. and Ross, P., 2002, Explorations in LCS models of stock trading, International Workshop on Learning Classifier Systems (IWLCS), 197-208.
[30]Simon, D. and Wiggins, R., 2001, S&P futures and contrary sentiment indicators, The Journal of Futures Markets, 447-462.
[31]Shimodaira, H., 1996, A New Genetic Algorithm Using Large Mutation Rates and Population-Elitist Selection (GALME), Proc.8th IEEE Conf. on Tools with Artificial Intelligence, 25-32.

[32]Syswerda, G., 1989, Uniform crossover in genetic algorithms, Proceeding of the 3rd International Conference on Genetic Algorithms, 2-9.
[33]Trippi, R. R., Sieno, D. D., 1992, Trading equity index futures with a neural network, Journal of Portfolio management, 27-33.
[34]Tuncay, C. and Stauffer, D., 2007, Power laws and Gaussians for stock market fluctuations, Physica A, 325–330.
[35]Xipin, C., Min, J., Bin, X., and Jie, C., 2000, Application of Elitist preserved Genetic Algorithms on Fuzzy controller, Proceedings of the 3th World Congress on Intelligent Control and Automation, Hefei, P.R.China.
[36]Yager, R., and Filev, D., 1994, Template-based fuzzy system modeling, Intelligent and Fuzzy System, Vol. 2, 39-54.
[37]William, L., Russell, P., and James, M. R., 2002, Forecasting the NYSE composite index with technical analysis, pattern recognizer, neural network, and genetic algorithm: a case study in romantic decision support, Decision Support Systems 32, 361– 377.