Using Artificial Neural Networks to Automatically Construct Rule Base for Forecasting Taiwan Electronic Companies' Stock Return and ROE Performance

植基於類神經網路之自動化規則庫建構應用於台灣電子公司股票報酬率暨股東權益報酬率預測之研究

10.6545/JFS.2009.17(1).6

左杰官(Brandt Tsoand)；簡旭生(Shad S. Jiang)

類神經網路 ； 股票報酬率 ； 股東權益報酬率 ； 專家系統 ； ANN ； stock retun ； ROE ； TREPAN ； expert system

財務金融學刊

17卷1期（2009 / 03 / 31）

173 - 195

英文

對投資者而言，股票報酬率（stock return）及股東權益級酬率（ROE）可用以瞭解投資報酬的程度及公司獲利狀況，因此，若能正確地對於該二項因素進行預測，對投資者將有莫大助益，有別於一般諮詢財務專家以進行預測的方式，本研究提出一種植基於類神經網路的方法，藉由對類神經網路進行解碼以能自動化建構規則庫並執行預測，所使用之網路為倒傳遞類神經網路，本研究引進TREPAN演算法以揭露隱匿於類神經網路內之知識，俾探究公司現行財務指標及下一季報酬狀況之關係，研究對象為台海股票市場上市電子公司，資料來源涵括公元2000至2005年。在本研究中僅考慮電子公司之基本面資訊，該等資訊相較於技術指標而言較難進行適當地關聯暨解釋，故而吾人希望透過此初步之研究能加速財務專家系統之規則庫建立，俾提供投資者更透明的軌跡暨資訊以增進對於公司狀況的診斷。實驗發現，利用基本面資訊作為輸入，類神經網路可達到約70.68%之正確在預測率，而本研究透過TREPAN演算法可成功地自類神經網路擷取相關規則，証明自動化建構財務規則庫之可行性，文中並揭示相關有趣的發現。

The stock returns and ROE are meaningful to the shareholders to realize the level of investment feedbacks and companies' profitability. The accurate forecasts for both factors thus can be very important to the investors Instead of consulting to the financial experts. this study proposes an approach by decoding artificial neural networks (ANN) to automatically construct a rule base for performing forecasts. The ANN being implemented is the so-called hack-propagntion neural network. The algorithm known as TREPAN is introduced to uncover the hidden knowledge from ANN for huilding the relationship between company's current financial indices and the probable performance in the next season. The study uses Taiwan stock market electronic companies in the time period from years 2000 to 2005 as a basis for carrying out experiments. The inputs for the ANN in this preliminary, study arc only concerned with the fundamental factors. It is expected that. through this empirical study, one may accelerate the rule base construction for the financial expert systems and to provide the more clear traces to improve the diagnosis to the companies. The results reveal that. using fundamental factors as inputs. the ANN can perform up to 70.68% accuracy in the experiments. In terms of TREPAN algorithm, the knowledge of companies' financial performance can be successfully extracted from ANN, though the minor error may occur. Some interesting discoveries are also addressed.

1. TSEC
2. Taiwan Economic Journal
3. Andrews, R.,S. Geva(1994).Rule extraction from a constrained error back-propagation multi-layer perceptron.Proc. of the Australian Conference on Neural Networks
4. Ariff, M.,T.K. Lim,M. Skully(2003).Accurate Prediction of Analyst Forecast Revisions and Stock Returns: EPS versus Non-EPS Variables.Global Business and Finance Review,8,61-75.
5. Browne, A.,B. D. Hudsonb,B. C. Whitley,M. G. Fordb,P. Pictonc(2004).Biological data mining with neural networks: implementation and application of a flexible decision tree extraction algorithm to genomic problem domains.Neurocomputing,57,275-293.
6. Chen, A. P.,M. Y. Chen(2006).Integrating extended classifier system and knowledge extraction model for financial investment predictions: An empirical study.Expert Systems with Applications,31,174-183.
7. Chen, A. P.,Y. C. Chen,Huang, U. H.(2005).Applying two-stage XCS model on global overnight effect for local stock prediction.Lecture Notes in AI,3681,34-40.
8. Chen, Q.,C. D. Li(2006).Comparison of Forecasting Performance of AR, STAR and ANN Models on the Chinese Stock Market Index.Proceedings of the 3rd International Symposium on Neural Networks,Chengdu, China:
9. Chen, Q.,I. Goldstein,W. Jiang(2003).Price Informativeness and Investment Sensitivity to Stock Price.The 14th Annual Conference on Financial Economics and Accounting (FEA),L.A., CA, USA.:
10. Craven, M. W.(1996).Madison, WI.,University of Wisconsin.
11. Deng, Z.,B. Lev,Narin, F.(1999).Science and Technology as Predictors of Stock Performance.Financial Analysts Journal,55,20-32.
12. Duch, W.,R. Adamczak,K. Grabczewski(2001).A new methodology of extraction, optimization and application of crisp and fuzzy logical rules.IEEE Transactions on Neural Networks,11,1-31.
13. Dunis, C. L.,J. Jalilov(2001).Neural Network Regression and Alternative Forecasting Techniques for Predicting Financial Variables.Neural Network world,12,113-139.
14. BBSRC Bioinformatics Grantholders Workshop
15. George, J. H.,P. Miller,R. Kerber(1996).Stock selection using rule induction.IEEE Transactions on Intelligent System,11,52-58.
16. Hüsken, M.,P. Stagge(2003).Recurrent Neural Networks for Time Series Classification.Neurocomputing,50,223-235.
17. Jasic, T.,D. Wood(2004).The profitability of daily stockmarket indices based on neural network predictions.Applied Financial Economics,14,285-297.
18. Jiang, S. S.(2006).Taiwan,National Defense Management College.
19. Kavajecz, K. A.,E. R. Odders-White(2004).Technical analysis and liquidity provision.Review of Financial Studies,17,1043-1071.
20. Lev, B.,R. Thiagarajan(1993).Fundamental Information Analysis.Journal of Accounting Research,31,190-215.
21. Milaré, C. R.,A. C. P. L. F. Carvalho,M. C. Monard(2002).An Approach to Explain Neural Networks using Symbolic Algorithms.International Journal of Computational Intelligence and Applications,2,365-376.
22. Morgan, S. P.,J. D. Teachman(1988).Logistic regression: description, examples, and comparisons.Journal of Marriage and the Family,50,925-936.
23. Murthi, B.,Y. Choi,P. Desai(1997).Efficiency of mutual funds and portfolio performance measurement: a nonparametric measurement.European Journal of Operational Research,98,408-418.
24. Pesaran, M. H.,A. Timmermann(2002).Market timing and return prediction under model instability.Journal of Empirical Finance,9,495-510.
25. Richardson, S. A.,R. G. Sloan,Rodney L.(2003).White Center for Financial Research Working Paper No. 03-03White Center for Financial Research Working Paper No. 03-03,未出版
26. Rumelhart, D. E.,G. E. Hinton,R. J. Williams(1986).Learning representations by back-propagating errors.Nature,323,533-536.
27. Schmitz, G. P. J.,C. Aldrich,F. S. Gouws(1999).ANN-DT: An Algorithm for extraction of decision trees from artificial neural networks.IEEE Transactions on Neural Networks,10,1392-1401.
28. Taha, I. A.,J. Ghosh(1999).Symbolic interpretation of artificial neural networks.IEEE Transactions on Neural Networks,11,448-463.
29. Towell, G.,J. W. Shavlik(1993).The extraction of refined rules from knowledge baaed neural networks.Machine Learning,31,71-101.
30. Tso, B.,P. M. Mather(2001).Classification methods for remotely sensed data.England U.K.:Taylor and Francis Ltd..
31. Wang, V.(2006).Yale School of Management.
32. Zekic, M.(1998).Croatia,University of Osijek.

1. 盧嘉梧、林志軒(2016)。類神經網路投資組合策略績效之實證研究：以台灣中型100電子股為例。輔仁管理評論，23(3)，29-50。