:::

詳目顯示

回上一頁
題名:應用事件分類元於金融事件之行為探勘 -以台灣股票除權息為例
作者:黃建華
作者(外文):Huang, Chien-Hua
校院名稱:國立交通大學
系所名稱:資訊管理研究所
指導教授:陳安斌
學位類別:博士
出版日期:2014
主題關鍵詞:事件分類元系統金融事件人工智慧分類元系統除權息波動率Event Classifier SystemFinancial EventsArtificial IntelligenceClassifier SystemEx-DividendVolatility
原始連結:連回原系統網址new window
相關次數:
  • 被引用次數被引用次數:期刊(0) 博士論文(0) 專書(0) 專書論文(0)
  • 排除自我引用排除自我引用:0
  • 共同引用共同引用:0
  • 點閱點閱:13
上市櫃公司的股價是來自於市場投資參與者的供需結果,基於行為財務學的發展,許多人工智慧的方法學被用來分析及預測股價下一步的變化,並挖掘出能有顯著獲利的交易規則,作為投資時機點的決策。但在交易的過程中,市場有各種影響投資人行為的資訊揭露及事件,這種金融事件影響的深度(漲跌幅度)與廣度(影響時間),在過去比較少被探討。本研究是以人工智慧方法學中的延伸性分類元系統(eXtended Classifier System, XCS)為基礎,透過知識探勘歸類、演化及評估來分析股票市場除權息事件之因果關係,探討事件發生時的股價行為是否具有預測性。
首先針對股票除權息事件歸納出相關的事件影響因子,並以台灣上市公司為樣本,蒐集2001年至2013年間的除權息事件的案例,建構以此事件為基礎的延伸性分類元系統,發展一套事件分類元系統(Event Classifier System, ECS)之學習及評估模式,探索影響因子在樣本內學習的影響程度,以及對於樣本外預測的準確性,以進行除權息事件的分析與探討。
實驗結果顯示,此除權息事件五種因子在樣本內的學習正確率能有90%以上的預測能力,對樣本外的測試亦能有70%左右的正確率,故在樣本內和樣本外的數據都優於事件隨機挑選的的正確率平均數56%,歸納各影響因子對於預測結果的分析,則以股票在除權息前的波動率、除權息折價率及除權前的股價走勢等三個方面影響較大,亦可以理解投資人在除權前的交易行為會反映除權後的股價走勢,說明事件交易的可預測性。最後,應用事件分類元系統作為投資決策的選擇,可以創造更佳的風險投資績效。
The stock price of listed companies is usually comes from the supply/demand of financial market investors, base on the development of behavioral finance, many of the artificial intelligence methodologies are used to analyze/forecast the changes in stock price. However, the behaviors of an investor will always be influenced by market events, the influence of rise/fall percentage and influence on time duration from financial events has less been discussed in the past yet. This study is done on the basis of eXtended Classifier System, through the method of valuation and data mining categorization to analyze the casual relationship between stock market behavior and ex-dividend event, in addition to investigate if stock price behaviors possess any predictability when an event occurs.
First of all, focus on ex-dividend events to categorize relative event influential factors, collecting ex-dividend/right events of Taiwan Public Listed Companies during the year of 2001 and 2003 as case study. Constructing an eXtended Classifier System based on the ex-dividend/right events, as a result an Event Classifier System - learning and valuation model has been created; exploring the significant of influence when an influential factor learn within the sample population, also the accuracy of predictability on data out of the original sample.
The result has shown, the five factors within the ex-dividend/right event possess higher than 90% accurate predictability when it learns within the sample population and approximately 70% accuracy when it predicts for out of the original sample. Due to the accuracy of statistical data collected in and out of the sample population are both surpassing if the data is to be collected randomly (by an average of 56% correctness). We can conclude the following factors as most influential: volatility of stock price before ex-dividend day, discounted price on ex-dividend day and stock trends before ex-dividend day. Hence, we can understand the trading behaviors of investors before ex-dividend day will have an influence on the stock trend after the ex-dividend day; illustrate the predictability of an financial event trading.Finally, Using Event Classifier System can create more stable performance of investment under the same risk.
[1] Barberis, Nicholas, Ming Huang, Tano Santos. “Prospect theory and asset prices”, Quarterly Journal of Economics 116, pp.1-53, 2001.
[2] Chen, A.P., Chen, Y.C., Tseng, W.C. “Applying Extending Classifier System to Develop an Option-Operation Suggestion Model of Intraday Trading-An Example of Taiwan Index Option”, Lecture Notes in AI, Vol.3681, pp.27-33, 2005.
[3] Chen, A.P., Chen, Y.C., Huang, Y.H. “Applying Two-Stage XCS Model on Global Overnight Effect for Local Stock Prediction”, Lecture Notes in AI, Vol.3681, pp.34-40, 2005.
[4] Fama, E.F. “Efficient Capital Market:A View of Theory and Empirical Work”, Journal of Finance, Vol.25, pp.383-417, 1970.
[5] Holland, J.H. “Processing and processors for Schemata”, Associative information processing, pp.127-146, 1971.
[6] Holland, J.H. “Progress in theoretical biology”, Adaptation. In R. Rosen &; F. M. Snell(Eds.), pp.263-293, 1976.
[7] Holland, J.H., Reitman, J.S. “Cognitive Systems Based on Adaptive Algorithms”, Pattern-Directed Inference Systems, Waterman D.A. &; Hayes-Roth F.(eds), Academic Press, New York, 1978.
[8] Holland, J.H. “Adaptive Algorithms for Discovering and using General Patterns in Growing Knowledge Bases”, International Journal of licy Analysis and Information Systems, 4(3), pp.245-268, 1980.
[9] Holland, J.H. “A Mathematical Framework for Studying Learning in Classifier Systems”, Physica, 2(1-3), pp.307-317, 1986.
[10] Kahneman, Daniel, Tversky Amos. “Prospect theory:An analysis of decision under risk”, Econometrica, Vol.47(2), pp.263-291, March 1979.
[11] Kahneman, Daniel, Riepe, Mark W. “Aspects of investor psychology”, Journal of Portfolio Management, pp.52-65, summer 1998.
[12] Lanzi, P.L., Riolo, R.L. “A Roadmap to the Last Decade of Learning Classifier System Research (From 1989 to 1999)”, Learning Classifier Systems:from Foundations to Applications, 1813, pp.33-62, 2000.
[13] Bull, L. Applications of Learning Classifier Systems, Springer-Verlag, Heidelberg, 2004
[14] Tversky, A., Kahneman, D. “Availability:a heuristic for judging frequency and probability”, Cognitive Psychology, 5, pp.207-232, 1973.
[15] Shefrin, H. Beyond Greed and Fear:Understanding Behavioral Finance and the Psychology of Investing, Massachusetts:Harvard Business School Press, Boston, 2000.
[16] Shleifer, A. Inefficient Market, Oxford U. Press, Oxford , 2000.
[17] Shiller, R. J. “Stock prices and social dynamics”, Brookings Papers on Economic Activity II, pp.457-498, 1984.
[18] Wilson, S.W., Goldberg, D.E. “A Critical Review of Classifier Systems”, In Proceedings of the 3rd International Conference on Genetic Algorithms, pp.244-255, 1989.
[19] Wilson, S.W. “ZCS: A zeroth level classifier system”, Evolutionary Computation, 2(1), pp.1-18, 1994.
[20] Wilson, S.W. “Classifier Fitness Based on Accuracy”, Evolutionary Computation, 3(2), pp.149-175, 1995.
[21] 鍾惠民、吳壽山、周賓鳳、范懷文,財金計量,台北,雙葉書局,民國91年。
[22] 周賓凰、池祥萱、周冠男、龔怡霖,『行為財務學:文獻回顧與展望』,證券市場發展季刊,第14卷第2期,1-47頁,民國91年。new window
[23] 曾文娟,『應用分類元股票交易系統於台灣加權指數趨勢預測之研究』,國立交通大學資訊管理研究所,碩士論文,民國93年。
[24] 蔡毓耕,『應用分類元系統於財務危機預警之研究』,國立交通大學資訊管理研究所,碩士論文,民國93年。
[25] 張永華,『基於反向情緒指標使用分類元系統預測美國S&;P期貨指數』,國立交通大學資訊管理研究所,碩士論文,民國93年。
[26] 陳安斌,新金融實驗教學之財務金融資訊系統與投資管理,初版,台北市,新陸書局股份有限公司,民國94年。
[27] 林昶立,『實數分類元系統於股票型基金投資組合建構之運用』,國立交通大學資訊管理研究所,碩士論文,民國94年。
[28] 楊雅君,『應用延伸性分類元系統預測利率期貨價格走勢-以十年期公債期貨為例』, 國立交通大學資訊管理研究所,碩士論文,民國95年。
[29] 傅世專,『應用分類元系統於台灣的指數期貨價差交易之實證研究』,國立交通大學資訊管理研究所,碩士論文,民國95年。
[30] 劉均偉、夏祥銓,“事件策略值得期待”,申銀萬國2013年冬季報告,第10章,1-22頁,2013。

 
 
 
 
第一頁 上一頁 下一頁 最後一頁 top
:::
無相關書籍
 
無相關著作
 
無相關點閱
 
QR Code
QRCODE