股票市場在現代資本市場上扮演舉足輕重的角色，所以吸引了相當多不同領域的人士對於金融商品預測的興趣。此外，在實務界被廣為接受的觀念之一是所謂股價的變化會依循某一個趨勢。因此，對於趨勢的判斷也成為股市預測中一項重要的任務。故此，本研究傾向預測長期的趨勢而非短期或單日內的價格變化。
雖然已有不少相關應用研究致力於趨勢的分析預測，但判斷趨勢的主要方法，多數都還是使用某一明確的門檻值來判斷不同的趨勢狀態。所以使用此模式的人就只知道判斷的結果是某一狀態，卻未被告知此建議的準確度或信心水準，從而進行決策時便缺少進一歨的訊息支援。由上述推論，本研究的目的即是利用動態貝氏網路，結合技術分析領域的鼻祖「道氏理論」（長期趨勢的預測原則），來試圖建構出一個具有預測能力的模式，以輔助投資者的長期投資決策。
透過本研究最後的實驗比較可得知，在觀察時間足以涵蓋該投資標的的完整趨勢周期時，本方法的投資報酬率具有擊敗買入持有投資策略的能力。這也代表著對於長期投資者而言，本方法具有獲得超額報酬的潛力，同時也能有效降低交易頻率與相對的交易成本。

Stock market plays an important role in the modern capital market. As a result, the prediction of financial assets attracts people in different areas. Moreover, it is commonly accepted that stock price movement generally follows a major trend. As a result, forecasting the market trend becomes an important mission for a prediction method. Accordingly, we will predict the long term trend rather than the movement of near future or change in a trading day as the target of our predicting approach.
Although there are various kinds of analyses for trend prediction, most of them use clear cuts or certain thresholds to classify the trends. Users (or investors) are not informed with the degrees of confidence associated with the recommendation or the trading signal. Therefore, in this research, we would like to study an approach that could offer the confidence of the trend analysis by providing the probabilities of each possible state given its historical data through Dynamic Bayesian Network. We will incorporate the well-known principles of Dow’s Theory to better model the trend of stock movements.
Through the results of our experiment, we may say that the financial performance of the proposed model is able to defeat the buy and hold trading strategy when the time scope covers the entire cycle of a trend. It also means that for the long term investors, our approach has high potential to win the excess return. At the same time, the trading frequency and correspondently trading costs can be reduced significantly.

1. Introduction 1
1.1 Background 1
1.2 Motivations and Objective 2
1.3 Problem Description－Efficient Market Hypothesis 3
2. Literature Review 6
2.1 Trend Analysis 6
2.2 Bayesian Network 8
2.3 Dynamic Bayesian Network 12
2.4 Related Work 15
2.5 Other Applications to Forecasting Stock Markets 16
3. Task Modeling 18
3.1 Stock market prediction 19
3.2 Concept model formation 21
3.3 Research process 25
4. Model Construction 26
4.1 Indicators study 28
4.2 Stock selection and data source 30
4.3 Data Preparations 32
4.4 Creating Model 47
4.5 Determining model configurations 53
5. Signaling strategy design and evaluation 56
5.1 Trials to improving signaling strategy 57
5.2 Strategy evaluation and results analysis 62
6. Conclusions 66
6.1 Findings and conclusion 67
6.2 Contribution 69
6.3 Limitation and further research opportunity 70
7. References 72
Appendix A Distributions of Trading Signals - 1 79
Appendix B Distributions of Trading Signals - 2 84

1.English literatures (sorted by alphabetic order)
[AS95] A. Saitta, “Using Dow Theory to Catch Trends,” Futures, October 1995, pp.46-48.
[DR83] David A. Glickstein, and Rolf E. Wubbels, “Dow Theory is Alive and Well,” Journal of Portfolio Management, Spring 1983, pp.28-32.
[EF91] E.F. Fama, “Efficient capital market: II”, Journal of Finance 46, p1575-1617, 1991.
[FGH97] N. Friedman, M. Goldszmidt, D. Heckerman, and S. Russell, “Challenge: Where is the impact of Bayesian networks in learning?”, In Proceedings of the Fifteenth International Joint Conference on Artificial Intelligence, Morgan Kaufmann Publishers, 1997.
[FT01] F. E. H. Tay et al, “Application of support vector machines in financial time series forecasting”, Omega International Journal of Management Science, 2001.
[GJ91] G. Leitch, J.E. Tanner, “Economic forecast evaluation: profits versus the conventional error measures”, American Economic Review (81): 580– 590, 1991.
[HB77] H.I Bishara, “Establishing A Modified Dow Theory for Canada and Testing Its Buy and Sell Signals Against A Buy and Hold Strategy,” Akron Business and Economic Review, pp.43-51, Winter 1977.
[HL00] H. Hong, T. Lim, J. Stein, “Bad news travels slowly: size, analyst coverage, and the profitability of momentum strategies”, The Journal of Finance LV (1): 265– 295, 2000.
[HMW95] D. Heckman, A. Mamdani, and M. Wellman, “Real-world applications of Bayesian Networks”, Communications of ACM, 38(3): 24-26, 1995.
[HP95] D. Heckman and M. Wellman, “Bayesian Networks”, Communications of ACM, 38(3): 27-30, 1995.
[HS99] H. Hong, J. Stein, “A unified theory of under-reaction, momentum trading, and overreaction in asset markets”, The Journal of Finance LIV (6) (1999) 2143– 2184.
[HW88] H. White, “Economic prediction using neural networks: The case of IBM stock daily returns”, In IEEE International Conference on Neural Networks, San Diego, P451-459, San Diego, 1988.
[JO04] Jangmin O et al, “Stock Trading by Modeling Price Trend with Dynamic Bayesian Networks”, In Proceedings of IDEAL 2004, 2004.
[JP82] J. Pear, “Reverend Bayes on inference engines: A distributed hierarchical approach”, In Proceedings AAAI National Conference on AI, pp.133-136, Pittsburgh, Pa, 1982.
[JP87] J. Pearl, “Evidential reasoning using stochastic simulation of causal models”, Artificial Intelligence, 32(2): 245-258, 1987.
[JP88] J. Pearl, “Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference”, San Mateo, CA: Morgan Kaufmann, 1988.
[JS00] J. Pearl and S. Russell, “Bayesian Networks”, Technical report, 2000.
[KK03] K. Kim, “Financial time series forecasting using support vector machines”, Neurocomputing, 55:307-319, 2003.
[KM98] K. P. Murphy, “A Brief Introduction to Graphical Models and Bayesian Networks”, http://www.ai.mit.edu/~murphyk/Bayes/bayes.html, 1998.
[KM01] K. P. Murphy, “The Bayes Net Toolbox for Matlab”, Computing Science and Statistics”, vol 33, 2001.
[KP83] J.H. Kim and J. Pearl, “A computational model for combined causal and diagnostic reasoning in inference systems.”, In Proceedings of IJCAI-83, pp.190-193, Karlsruhe, Germany, 1983.
[LS04] L. Shen et al, “Applying Rough Set to market timing decision”, Decision Support System, 2004.
[LS88] S. L. Lauritzen and D. J. Spiegelhalter, "Local computations with probabilities on graphical structure and applications to expert systems", J. Royal Stat. Soc. B, 50(2), 1988.
[LT91] G. Leitch and J.E. Tanner, “Economic forecast evaluation: profits versus the conventional error measures”, American Economic Review 81: 580– 590, 1991.
[LW91] D. Lowe and A.R. Webb. “Time series prediction by adaptive networks: A dynamic systems perspective.” IEEE Computer Society Press, 1991.
[MG04] M.J.A. Berry and G.S. Linoff, “Data Mining Techniques for Marketing, Sales and Customer Relationship Management”, Wiley Publishing Inc., Indianapolis, Indiana, 2004.
[MJ98] M.I. Jordan et al, “An introduction to variational methods for graphical models”, In M.I. Jordan, editor, Learning in Graphical Models. Kluwer, Dordrecht, The Netherlands, 1998.
[MK02] M. Ausloos and K. Ivanova, “Mechanistic approach to generalized technical analysis of share prices and stock market indices”, The European Physical Journal B 27: 177– 187, 2002.
[MP01] K. Murphy and M. Paskin, “Linear time inference in hierarchical HMMs”, Proceedings of Neural Information Processing Systems, 2001.
[MP91] M. J. Pring, “Technical analysis explained - The Successful Investor's Guide to Spotting Investment Trends and Turning Points”, New York McGraw-Hill Professional, 1991.
[MS03] M. Scott, “Charles Dow’s Six Secrets to Market Success, Applying the Dow Theory to Today’s Markets”, Alchemist issue 30, 2003.
[ND98] N. Drakos, “From A General Architecture for Supervised Classification of Multivariate Time Series”, technical report, 1998.
[PS98] P. Smyth, "Belief networks, hidden Markov models, and Markov random fields: a unifying view", Pattern Recognition Letters, 1998.
[RD99] P. Raghubir and S. R. Das, “A case for theory-driven experimental enquiry”, Financial Analysts Journal, November/December of 1999, pp.56–79, 1999.
[RN95] S. Russell, P. Norvig, “ Artificial Intelligence: A Modern Approach”, 2nd edition, Prentice Hall Series in Artificial Intelligence, Englewood Cliffs, New Jersey, 1995.
[SB98] S.J. Brown, W. A. Goetzmann, and A. Kumar, “The Dow Theory: William Peter Hamilton’s Track Record Reconsidered,” The Journal of Finance, Auguest 1998,pp.1311-1333.
[SL95] S.L. Lauritzen, “The EM algorithm for graphical association models with missing data”, Computational Statistics and Data Analysis, 19: 191-201, 1995.
[ST00] L. Shen, E.H. Tay, “Classifying market states with WARS”, Proceedings of the 2nd International Conference on Data Mining, Financial Engineering, and Intelligent Agents (IDEAL 2000), Springer, New York, pp. 280–285, 2000.
[TS02] E.H. Tay, L. Shen, “A modified Chi2 algorithm for discretization”, IEEE Transactions on Knowledge and Data Engineering, 14(3): 666– 670, 2002.
[Web-1] K.P.Murphy, “How to use the Bayes Net Toolbox”, http://www.cs.ubc.ca/~murphyk/Software/BNT/usage.html.
[Web-2] K.P.Murphy, “How to use BNT for DBNs”, http://www.cs.ubc.ca/~murphyk/Software/BNT/usage_dbn.html.
[WH1922] William P. Hamilton, “The Stock Market Barometer: A Study of its Forecast Value Based on Charles H. Dow’s Theory of the Price Movement”, Barron’s, New York, NY, 1922.
[WL02] W. Leigh et al, “Forecasting the NYSE composite index with technical analysis, pattern recognizer, neural network, and genetic algorithm”, Decision support system, Vol. 32, 2002.
[XW02] X. Wang and Z. Wang, “Stock market time series data mining based on regularized neural network and rough set”, Proceeding of International Conference of Machine Learning and Cybernetics 2002, Vol. 1, P315~318, 2002.
[ZP96] N.L. Zhang and D. Poole, “Exploiting causal independence in Bayesian network inference”, Journal of Artificial Intelligence Research, 5: 301-328, 1996.

2.Chinese Literatures
[王盈傑04] 王盈傑,”台灣股市動量策略與反向策略之整合研究”, 中正經濟研究所, 2004.
[李怡恬04] 李怡恬,”效率市場假說之實證研究－以亞太地區股市為例”, 銘傳管科所, 2004.
[陳寶妃04] 陳寶妃, ”台灣股市節慶效應之實證研究”, 實踐企管所, 2004.
[黃怡芬01] 黃怡芬, ”道氏理論、濾嘴法則與買入持有策略在台灣股市投資績效之比較”, 成功大學企業管理研究所碩士論文, 2001.
[CWeb-1] 參考網頁http://www.warrant.idv.tw/teach/learn.asp, ”技術分析教學”