臺灣博碩士論文加值系統

財務時間序列預測向來是投資決策重要的議題，找出良好的預測模式在時間序列預測研究中向來被視為極具挑戰性的應用領域。早期的研究大都採用統計方法假設，近年乃以人工智慧為基礎之各種演算法，例如支援向量機(support vector machine, SVM)等。本研究嘗試應用一新發展且在其他研究領域有良好預估績效但卻較少應用於財務時間序列研究的人工智慧預測方法-隨機森林(random forests, RF)。此外在財務時間序列資料中，因具有雜訊性質，且為驗證RF的可行性及提升其預測績效，本研究提出一結合獨立成分分析(independent component analysis)與隨機森林之二階段模式建構程序。首先將預測變數利用獨立成分分析，得到去除雜訊後之預測變數後，再使用RF以去除雜訊後之預測變數建構預測模式。期望以隨機森林進行預測時不受雜訊之影響，進而提升預測結果的準確度。為驗證所提方法之有效性，本研究以S&P500收盤指數之漲跌變化進行實證研究，並與支援向量機之預估結果進行比較。實證結果顯示，所提之方法之預測績效較支援向量機為佳。

To construct the better in prediction model is one of the most challenging ap-plications of modern financial time series forecasting. In terms of forecasting ap-proaches, early researches relied on conventional statistic methods, while recent stu-dies tended to apply artificial intelligence based techniques, such as support vector machine (SVM). The main objective of this research is to propose a forecasting tech-nique for the financial time series forecasting based on random forest (RF), a novel forecasting technique famous for artificial intelligence. As financial time series are inherently noisy, a two-stage approach by integrating independent component analy-sis (ICA) and RF is proposed in this study for financial time series forecasting. The proposed approach first uses ICA to the forecasting variables for generating the in-dependent components (ICs).
After identifying and removing the ICs containing the noise, the rest of the ICs are then used to reconstruct the forecasting variables which contain less noise. The RF is then applied to use the filtered forecasting variables to build the forecasting model. In order to evaluate the performance of the proposed approach, the Standard and Poor’s500 index is used as the illustrative example. A RF based forecasting model and SVM model were selected as the benchmark. The result shows that are expected to greatly expand the application of the proposed approach to forecast the stock change, and better business modeling and investment decisions can be found and implemented.

中文摘要　．．．．．．．．．．．．．．．．．．．．．．iii
英文摘要　．．．．．．．．．．．．．．．．．．．．．．iv
誌謝辭　　．．．．．．．．．．．．．．．．．．．．．．v
內容目錄　．．．．．．．．．．．．．．．．．．．．．．vi
表目錄　　．．．．．．．．．．．．．．．．．．．．．．viii
圖目錄　　．．．．．．．．．．．．．．．．．．．．．．ix
第一章　　緒論．．．．．．．．．．．．．．．．．．．．1
　　第一節　　研究背景與動機．．．．．．．．．．．．．1
　　第二節　　研究目的．．．．．．．．．．．．．．．．3
第三節　　論文架構．．．．．．．．．．．．．．．．．．4
第二章　　文獻探討．．．．．．．．．．．．．．．．．．6
　　第一節　　股市漲跌的意義與財務時間序列預測．．．．6
　　第二節　　隨機森林．．．．．．．．．．．．．．．．9
　 第三節　　獨立成份分析．．．．．．．．．．．．．．10
第四節 支援向量機．．．．．．．．．．．．．．．13
第三章　　研究方法．．．．．．．．．．．．．．．．．．16
　　第一節　　獨立成份分析．．．．．．．．．．．．．．16
　　第二節　　隨機森林．．．．．．．．．．．．．．．．18
第三節　　支援向量機．．．．．．．．．．．．．．．．．23
第四章　　實證與結果分析．．．．．．．．．．．．．．．26
　　第一節　　研究設計與資料收集．．．．．．．．．．．25
第二節　　獨立成份分析．．．．．．．．．．．．．．．．29
第三節 隨機森林．．．．．．．．．．．．．．．．．．31
第四節 支援向量機．．．．．．．．．．．．．．．．．32
第五節 比較分析結果．．．．．．．．．．．．．．．．33
第五章　　結論與建議．．．．．．．．．．．．．．．．．35
　　第一節　　研究結論．．．．．．．．．．．．．．．．35
　　第二節　　研究貢獻與建議．．．．．．．．．．．．．36
參考文獻　．．．．．．．．．．．．．．．．．．．．．．37

參考文獻
一、中文部分

呂奇傑，李天行，高人龍，陳學群(2008)，結合獨立成份分析與類神經網路於時間序列預測模式之建構，Chiao Da Management Review，28(2)，187-216。

連立川，葉怡成(2008)，以遺傳神經網路建構台灣股市買賣決策系統之研究，資訊管理學報，15(1)，29-51。

陳安斌，許育嘉(2004)，整合小波轉換與類神經網路於金融投資決策時間序列之研究，資訊管理學報，11(1)，139-165。

黃宇翔，王百祿(2008)，ARIMA與適應性SVM之混合模型於股價指數預測之應用，Journal of e-Business，10(4)，1041-1066。









二、英文部分

Allen, F., & Karlajainen, R. (1999). Using genetic algorithms to find technical trading rules. Journal of Financial Economics 51(2), 245-271.

Balachandher, K. G., Fauzias, M. N., & Lai, M. M. (2002). An ex-amination of the random walk model and technique trading rules in the Malaysian to stock market. Quarterly Journal of Business & Economics, 41(3), 81-104.

Bart, L., & Dirk, V. D. P. (2005). Predicting customer retention and profitability by using random forests and regression forests techniques. Expert Systems with Application, 29(1), 472-484.

Breiman, L. (1996). Bagging predictors. Machine Learning, 24(2), 123-140.

Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5-32.

Brok, W. A., & Kleidon, A. W. (1992). Periodic market closure and trading volume: A model of intraday bids and asks. Journal of Economic Dynamics and Control, 16(3-4), 451-489.

Buckinx, W., & Dirk, V. D. P. (2005). Customer base analysis: Partial defection of behaviourally-loyal clients in a non-contractual FMCG retail setting. European Journal of Operational Research, 164(1), 252-268.

Cao, L., & Gu, Q. (2002). Dynamic support vector machines for non-stationary time series forecasting. Intelligent Data Anal-ysis, 6(1), 67-83.

Chang, C. C., & Lin, C. J. (2001). LIBSVM：A library for support vector machines [Online]. Available: http://www.csie.ntu.ed u.tw/~cjlin/libsvm/ [2008, May 8]

Fama, E. F. (1970). Efficient capital market: A review of theory and emprical. Journal of Finance, 25(2), 383-417.

Harb. R., Xuedong, Y., Essam, R., & Xiaogang, S. (2009). Exploring precrash maneuvers using classification trees and random forests. Accident Analysis and Prevention, 41(1), 98-107.

Huang, W., Nakamori, Y., & Wang, S. Y. (2005). Forecasting stock market movement direction with support vector machine. Computer and Operations Research, 32(10), 2513-2522

Kim, K. J. (2003). Financial time series forecasting using support vector machines. Neurocomputing, 55(1-2), 307-319.

Kiviluoto, K., & Oja, E. (1998, October 21-23). Independent component analysis for parallel financial time series. Paper presented at proceedings of the Fifth International Conference on Neural Information, Kitakyushu, Japan.

Leigh, W., Hightower, R., & Modani, N. (2005). Forecasting the New York stock exchange composite index with past price and interest rate on condition of volume spike. Expert systems with Applications, 28(1), 1-8

Liaw, A., & Wiener, M. (2002). Classification and regression by random forest. R News, 2(3) , 18-22.

Lu, C. J., Lee, T. S., & Chiu, C. C., (2009). Financial time series forecasting using independent component analysis and support vector regression. Decision Support Systems, 47(2), 115-125.

Oja, E., Kiviluoto, K., & Malaroiu, S. (2000) Independent component analysis for financial time series, in Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing. Communications and Control Symposium, 111-116.

Pardo, M., & Sberveglieri, M. (2008). Random forests and nearest shrunken centroids for the classification of sensor array data. Sensors and Actuators B: Chemical, 131(1), 93-99.

Sánchez, V. D. A. (2003). Advanced support vector machines and kernel methods. Neurocomputing, 55(1), 5-20.

Tay, F. E. H., & Cao, L. J. (2001). Application of support vector machines in financial time series forecasting. Omega, 29(4), 309-317.

Tay, F. E. H., & Cao, L. J. (2002). Modified support vector ma-chines in financial time series forecasting. Neurocomputing, 48(5), 847-861.

Visser, E., & Lee, T. W. (2003), Speech enhancement using blind source separation and two-channel energy based speaker detection, Proceedings of 2003 IEEE International Conference on Acoustics, Speech, and Signal, 884-887.

Yaser, S., & Atiya, A. F. (1996). Introduction to financial forecasting. Applied Intelligence, 6(3), 2-7.

Xie, Y., Li, X., Nagi, E. W .T., & Ying, W. (2009). Customer churn prediction using improved balanced random forests. Expert Systems with applications, 36(3), 5445-5449.

Yim, J. (2002). A comparison of neural networks with time series models for forecasting returns on a stock market index. Lecture Notes in Computer Science, 58(6), 25-35.

Zheng, L., Watson, D. G., Johnston, B. F., Clark, R. L., Edrada, E. R., & Elseheri, W. (2008). A chemo metric study of chromatograms of tea extracts by correlation optimization warping in conjunction with PCA, support vector machines, and random forest data modeling. Analytica Chimica Acta, 642(1-2) , 257
- 265.