隨著台灣經濟的快速發展，投資理財已非常盛行，如股票和期貨已成為一般投資者獲利的主要投資標的之一。若能準確地預測股票或期貨之價格，必能協助投資者達成更正確之投資判斷與決策，以獲取更高的投資報酬。因此，對於股票或期貨之價格預測已成為熱門之財務研究相關議題，也吸引了很多來自實務投資者或學術研究者的注意。故本研究利用因素分析、基因演算法、階層群集分析、k-menas分群法和遺傳程式規劃，發展一個股票/期貨指數的系統化預測程序。其中，因素分析是用以協助研究者在許多技術指標中挑選對未來股票/期貨指數影響較為顯著之指標，而基因演算法、階層群集分析、k-menas分群法則用以將資料進行分群，以期望對於各群集之資料能夠建立更精確之預測模型，最後，則利用遺傳程式規劃法針對各群集之資料建構股票/期貨指數預測模型。同時，本研究以預測2001年3月1日至2010年3月31日的台灣指數期貨之隔日收盤價為例，驗證本研究所提出預測程序之可行性與有效性。實證結果顯示，本研究所提出之預測程序可以非常精確地預測台灣指數期貨之隔日收盤價，且比較結果亦顯示本研究所提出預測程序之預測準確度明顯優於單純只透過遺傳程式規劃或是經過因素分析進行屬性篩選然後以遺傳程式規劃所建構之預測模型。因此，本研究認為我們所提出之系統化預測程序可以實際應於協助投資者預測未來之股票或期貨的隔日收盤價，以協助其制定更好之投資決策。

Financial investment is very popular in Taiwan as the economies developed rapidly. Stocks and futures have become the major investing instruments for investors to earn profits. It is the crucial issue of financial research to forecast the prices of stocks and futures. We adopted the factor analysis, genetic algorithms, hierarchical cluster analysis, k-means analysis and genetic programming to develop a systematic forecast procedure for stock/futures index. Factor analysis is used to select the main factors of stock/futures index. We adopt genetic algorithms, hierarchical cluster analysis and k-means analysis to classify the data for setting up the better forecast model. Finally, genetic programming is used to construct the forecast model. We test the forecast procedure by forecasting the next close price of Taiwan index futures from March 1, 2001 to March 31, 2010. The results show that the forecast procedure we developed is able to forecast the next close price of Taiwan index futures very exactly, better than those by the pure genetic programming model and the genetic programming model with feature selection by using factor analysis. That means our systematic forecast procedure can be applied to forecast the future close price of stock/futures and help investors to make a better portfolio strategy.

中文摘要 ......................................................i
英文摘要 .....................................................ii
誌謝 ........................................................iii
目錄 .........................................................iv
表目錄 .......................................................vi
圖目錄 .................................................... viii
第一章 緒論 .................................................. 1
1.1 研究背景與動機 ........................................... 1
1.2 研究目的 ................................................. 2
1.3 研究流程 ................................................. 2
第二章 文獻探討 .............................................. 4
2.1 期貨或股票相關之研究 ..................................... 4
2.2 基因演算法 ................................................7
2.3 遺傳程式規劃 .............................................12
2.4 集群分析 .................................................17
2.5 屬性篩選 .................................................21
第三章 研究方法 ..............................................25
3.1 研究流程與步驟 .......................................... 25
3.2 蒐集資料 .................................................26
3.3 計算技術指標 .............................................27
3.4 數據正規化 ...............................................31
3.5 屬性篩選 ................................................ 32
3.6 分群 .....................................................34
3.7 以遺傳程式規劃預測隔日期貨價格 ...........................38
3.8 結果與分析 ...............................................39
第四章 實例驗證與分析 ....................................... 40
v
4.1 資料蒐集 ................................................ 40
4.2 計算技術指標 .............................................40
4.3 數據正規劃 .............................................. 41
4.4 屬性篩選 ................................................ 41
4.5 資料分群 ................................................ 43
4.6 以遺傳程式規劃預測隔日期貨價格 .......................... 49
4.7 以其他程序預測台指期貨隔日收盤價 .........................60
4.8 比較與分析 ...............................................64
第五章 結論與建議 ........................................... 67
5.1 研究結論 ................................................ 67
5.2 管理意涵 ................................................ 68
5.3 研究限制 ................................................ 69
5.3 未來研究建議 ............................................ 70
附錄 ........................................................ 71
附錄A 台指期貨收盤價實際值與預測值之比較圖(以FS-GA-GP 為預測方法)...........................................................71
附錄B 台指期貨收盤價實際值與預測值之比較圖(以FS-HA-GP 為預測方法).......................................................... 73
附錄C 台指期貨收盤價實際值與預測值之比較圖(以FS-KM-GP 為預測方法)...........................................................77
附錄D 台指期貨收盤價實際值與預測值之比較圖(以FS-NCA-GP 為預測方法)...........................................................80
附錄E 台指期貨收盤價實際值與預測值之比較圖(以NFS-NCA-GP 為預測方法)...........................................................83

參考文獻 .................................................... 86

中文文獻
1. 古永嘉和許世璋，2009，應用狀態空間模型與基因類神經網路濾波技術於風
險值預測之研究：以台股指數與台指期貨為例，台大管理論叢，20(2)，
pp.307-342。
2. 杜金龍，2002，技術指標在台灣股市應用之訣竅，初版，臺北市 : 財訊出版 :
聯豐書報總經銷。
3. 辛永森，2008，台灣股價指數期貨預測－平滑支撐向量迴歸與灰預測之應用，
國立台灣科技大學資訊管理系。
4. 陳國玄，2004，人工神經網路與統計方法應用於台灣上市電子類股價指數預
測與分類之研究，國立成功大學統計學系碩博士班。
5. 陳界榕，2009，應用兩階段預測方式於台指期貨之研究，輔仁大學資訊管理
學系。
6. 陳適宜，2009，基因類神經網路在臺股指數期貨之預測與蝶式交易策略研究，
國立台北大學企業管理學系。
7. 陳子安，2010，薄型晶圓片切割參數最佳化知研究-以蕭特基二極體為例，明
新科技大學企業管理研究所。
8. 葉青，2010，利用粗糙集合動態縮減集支援股市買賣決策，南台科技大學國際企
業系。
9. 蔡美枝，2010，應用級比檢驗建立擇股策略整合決策模型，朝陽科技大學財務金
融系。
10. 蔡承益，2007，使用SOM-SVR 混合型系統搭配屬性篩選模式應用於台灣指數
期貨預測，國立高雄第一科技大學資訊管理所。
11. 鄭家豪，2007，灰關聯雙移動平均線之應用研究-以股票市場及金屬期貨為例，
國立成功大學資源工程學系博士班。
英文文獻
1. Akbar, E., and Somayeh M., 2011, A genetic programming model to generate
risk-adjusted technical trading rules in stock markets, Ecpert Systems with
Applications, 38(7), pp.8438-8445.
2. Chantziara, T., and Skiadopoulos G., 2008, Can the dynamics of the term structure
of petroleum futures be forecasted? Evidence from major markets, Energy
Economics, 30(3), pp.962-985.
3. Chen, C.H., Chen T.L., and Wei L.Y., 2010, A hybrid model based on rough sets
theory and gentic algorithms for stock price forecasting, Information Sciences,
180(9), pp.1610-1629.
4. Chen, T.S., Chen J., Liu C.H., and Guo Y.S., 2008, Hierarchical clustering
investment portfolio - An example using the Taiex 50 stocks, Book Series: Series
of Information and Management Sciences 7, pp.112-114.
5. Cuong, T. and Tuan D.P., 2009, Analysis of cardiac imaging data using decision
tree based parallel genetic programming, Proceedings of 6th International
Symposium on Image and Signal Processing and Analysis, pp.317-320.
6. Cinar, V., Oncan T., and Sural H., 2010, A genetic algorithm for the traveling
salesman problem with pickup and delivery using depot removal and insertion
moves, Lecture Notes in Computer Science, 60(25), pp.431-440.
7. Deng, Y., and Liu J., 2009, Feature selection based on mutual information for
language recognition, Proceedings of The 2009 2nd International Congress on
Image and Signal and signal processing, 1-9, pp.4319-4322.
8. Dunteman, G.H., 1994, Principal component analysis. In M. S. Lewis-Beek (Eds.),
Factor analysis and related techniques, pp.157-245. Sara Miller McCune, CA: Sage
Publications, Inc.
9. Fathi, H., and Afshar A., 2010, GA-based multi-objective optimization of
finance-based construction project scheduling, Ksce Journal of Civil Engineering,
14(5), pp.627-638.
10. Ghoseiri, K., and Ghannadpour S.F., 2010, Multi-obective vehicle routing problem
with time windows using goal programming and genetic algorithm, Applied Soft
Computing, 10(4), pp.1096-1107.
11. Holland, J. H., 1975, Adaptation in natural and artificial systems, Ann Arbor, MI:
The University of Michigan Press.
12. Huang, W.L., Chen L.Z., and Li G.J., 2002, Comprehensive application of factor
analysis and hierarchical cluster analysis in evaluating the effect of FDI,
Proceedings of 2002 International Conference on Management Science and
Engineering, pp.1270-1274.
13. Hsu, L.Y., Horng S.J., Kao T.W., Chen Y. H., Run R.S., Chen R.J., Lai J.L., and
Kuo I.H., 2010, Temperature prediction and TAIFEX forecasting based on fuzzy
relationships and MTPSO techniques, Expert Systems with Applications: An
International Journal, 37(4), pp.2756-2770.
14. Joseph, S., Sheriff and Ayers R., 2003, Intrusion detection: methods and system.
part II, Information Management and computer security, 11(5), pp. 222-229.
15. Koza, J. R., 1992, Genetic programming - on the programming of computers by
means of natural selection, Cambridge, MA. MIT Press.
16. Lee, H.S., Roh S., Park M.S., and Ryu H.G., 2010, Optimal option selection for
finishing works of high-rise building, KSCE Journal of Civil Engineering, 14(5),
pp.639-651.
17. Lin, S.W., Shiue Y.U., Chen S.C., and Cheng H.M., 2009, Applying enhanced data
mining approaches in predicting bank performance: A case of Taiwanese
commercial banks, Expert Systems with Applications, 36(9), pp.11543-11551.
18. Li, S.B., Pan W.J., Yang G.C., and Chen L.N., 2009, Optimization of 3G wireless
network using genetic programming, Second International Symposium on
computational intelligence and design, 2, pp.131-134.
19. Liu, H., Motoda H., 1998, Feature selection for knowledge discovery and data
mining. Boston: Kluwer Academic Publishers.
20. Spearman, C., 1904, General intelligence, Objectively determined and measured,
The American Journal of Psychology, 16(2), pp.201-293.
21. Pai, G.A.V., and Michel T., 2009, Evolutionary Optimization of Constrained
k-means Clustered Assets for Diversification in Small Portfolios, IEEE
Transactions on Evolutionary Computation, 13(5), pp.1030-1053.
22. Vasant, P., and Barsoum N., 2009, Hybrid simulated annealing and genetic
algorithms for industrial production management problems, International Journal
of Computational Methods, 7(2), pp.254-261.
23. Versace, M., Bhatt R., Hinds O., and Shiffer M., 2004, Predicting the exchange
traded fund DIA with a combination of genetic algorithms and neural networks,
Expert Systemswith Applications, 27(3), pp.417-425.
24. Xu, G.X., Shia B.C, Hsu Y.B., Shen P.C., and Chu K.H., 2009, To integrate text
mining and artificial neural network to forecast gold futures price, International
conference on new trends in information and service science, 1(2), pp.1014-1020.
25. Yang, G.P., Zhou G.T., Yin Y.L., and Yang X.K., 2010, K-means based
fingerprint segmentation with sensor interoperability, Eurasip Journal on
Advances in Signal Processing.
26. Zhao, X., Liu X., Hao X.Y., and Liu K.Y., 2009, An algorithm of feature selection
and feature weighting adjustment based on Chinese Frame Net, Proceedings of The
2009 2nd International Congress on Image and Signal and signal processing, 1-9,
pp.4300-4303.