臺灣博碩士論文加值系統

財務困境預測議題一直以來在財務金融領域佔有很重要且具有挑戰性的研究主題。目前已經有很多預測公司破產和金融危機方法，包括人工智慧和統計方法，研究結果顯示人工智慧之預測效果比傳統統計方法佳。由於公司的財務報表是季報表，所以影響公司財務困境的屬性資料是季節型時間序列，而財務困境資料具有不平衡類別且非穩定性的時間序列資料。本研究利用機器學習技術建立了兩個時間序列財務困境模型。在第一個模型中提出了一種規則化基因表示規劃法模型。 在第二個模型中提出了一個MetaCost在分類器的訓練中增加了成本敏感的分類。所提出的模型有幾個優點：（1）利用MetaCost算法處理不平衡類; （2）提出的模型是季節性時間序列模型; （3）採用屬性選擇來尋找核心屬性，降低維度資料量; （4）具有說服力的解釋能力與管理意涵，提供投資人與決策者參考。結果表明, 所提出的 GEP 和 Metacost 方法皆優於所列出的分類器, 本研究提出的人工智慧GEP 預測結果具有相對的準確度，第一類型錯誤與第二類型錯誤的優勢。Metacost方法提高了一些敏感性,利於確定實際財務健康的公司;第二型錯誤21.6%, 表明該方法可以提高財務困境的正確分類並且定義出屬性範圍值提供財務困境參考。研究結果可以使投資者及早發現公司的財務困境程度。

Financial distress prediction is an important and challenging issue in the financial field. Now, many methods have been proposed to forecast company bankruptcy and financial crisis, and many studies show that artificial intelligence is better than traditional statistical methods in prediction capacity. Financial statements are quarterly reports; hence the attributes of a financial crisis are quarterly time-series data, and financial crisis data have the properties of the imbalance class and are non-stationary. This study uses machine learning techniques to build two time-series financial distress model. In the first model proposed a gene expression programming model for predicting the financial distress of companies. In the second model proposes a MetaCost to add cost-sensitive classification in the training of base classifiers. The proposed model has the following advantages: (1) utilize the MetaCost algorithm to handle the imbalance class; (2) the proposed model is a time series model and is calculated in quarters.; (3) employ attribute selection to find the core attributes and reduce data dimension; (4) the results of the study can be used as a reference for investors and decision makers. The results show that the proposed GEP and Metacost methods are better than the listed classifiers. The prediction results of the artificial intelligence GEP method proposed in this study have relative accuracy, and the advantages of Type I error and Type II error. MetaCost method raises a little sensitivity, it lifts to identify the companies’ financial health when the companies are actually healthy; and type II errors are reduced by 21.6%, it denotes that the proposed method can raise the correct classification of financial distress and define the attribute range value to provide financial distress reference the results can early enable the investors to detect the level of financial distress of a corporation.

Table of contents
摘 要 i
Abstract ii
誌 謝 iii
Table of contents iv
List of tables vi
List of figures vii
1. Introduction 1
2. Literature review 4
2.1. Financial distress 4
2.2. MetaCost 16
2.3. Attribute selection 18
2.4. Gene Expression Programming (GEP) 22
2.5. Time series 23
2.6. Classification technology 24
2.6.1. Decision tree (DT) 25
2.6.2. Support vector machine (SVM) 25
2.6.3. Multilayer Perceptron (MLP) 26
2.6.4. Radial Basis Function (RBF) Network 29
2.6.5. K-Nearest Neighbor (KNN) 30
3. Proposed method 32
3.1. Research concept 32
3.2. Proposed Time-series financial distress prediction model 34
4. Experiments and Comparisons 44
4.1. Selecting Attributes 47
4.2. GEP model experiment 48
4.3. Moving Window Experiment 51
4.3.1. GEP moving window experiment 52
4.3.2. Metacost moving window experiment 52
4.4. Findings 58
4.4.1. Important attributes 58
4.5. Model application: 59
5. Conclusion 67
5.1. Limitation 69
5.2. Future work 70
References 71

Altman, E. I. (1968). Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. The journal of finance, 23(4), 589-609.
Altman, N. S. (1992). An introduction to kernel and nearest-neighbor nonparametric regression. The American Statistician, 46(3), 175-185.
Atiya, A. F. (2001). Bankruptcy prediction for credit risk using neural networks: A survey and new results. IEEE transactions on neural networks, 12(4), 929-935.
Auria, L., & Moro, R. A. (2008). Support vector machines (SVM) as a technique for solvency analysis.
Bauer, E., & Kohavi, R. (1999). An empirical comparison of voting classification algorithms: Bagging, boosting, and variants. Machine learning, 36(1-2), 105-139.
Beaver, W. H. (1966). Financial ratios as predictors of failure. Journal of accounting research, 71-111.
Bishop, C., & Bishop, C. M. (1995). Neural networks for pattern recognition: Oxford university press.
Borges, H. B., & Nievola, J. C. (2012). Comparing the dimensionality reduction methods in gene expression databases. Expert Systems with Applications, 39(12), 10780-10795.
Box, G., & Jenkins, G. (1970). Time Series Analysis Forecasting and Control/Holden Day, San Francisco, California. In.
Broomhead, D. S., & Lowe, D. (1988). Radial basis functions, multi-variable functional interpolation and adaptive networks. Retrieved from
Chen, M.-Y. (2011). Predicting corporate financial distress based on integration of decision tree classification and logistic regression. Expert Systems with Applications, 38(9), 11261-11272.
Chen, N., Ribeiro, B., Vieira, A. S., Duarte, J., & Neves, J. C. (2011). A genetic algorithm-based approach to cost-sensitive bankruptcy prediction. Expert Systems with Applications, 38(10), 12939-12945.
Chen, Y., Zhang, L., & Zhang, L. (2013). Financial distress prediction for Chinese listed manufacturing companies. Procedia Computer Science, 17, 678-686.
Cheng, C.-H., & Wang, S.-H. (2015). A quarterly time-series classifier based on a reduced-dimension generated rules method for identifying financial distress. Quantitative Finance, 15(12), 1979-1994.
Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20(3), 273-297.
Czajkowski, M., Czerwonka, M., & Kretowski, M. (2015). Cost-sensitive global model trees applied to loan charge-off forecasting. Decision Support Systems, 74, 57-66.
Dayan, P., & Berridge, K. C. (2014). Model-based and model-free Pavlovian reward learning: revaluation, revision, and revelation. Cognitive, Affective, & Behavioral Neuroscience, 14(2), 473-492.
Deakin, E. B. (1972). A discriminant analysis of predictors of business failure. Journal of accounting research, 167-179.
Domingos, P. (1999). Metacost: A general method for making classifiers cost-sensitive. Paper presented at the KDD.
Eiben, A. E., & Smith, J. E. (2003). Introduction to evolutionary computing (Vol. 53): Springer.
Fallahpour, S., Lakvan, E. N., & Zadeh, M. H. (2017). Using an ensemble classifier based on sequential floating forward selection for financial distress prediction problem. Journal of Retailing and Consumer Services, 34, 159-167.
Ferreira, C. (2001). Algorithm for solving gene expression programming: a new adaptive problems. Complex Systems, 13(2), 87-129.
Foster, G. (1978). Financial Statement Analysis, 2/e: Pearson Education India.
Frydman, H., Altman, E. I., & KAO, D. L. (1985). Introducing recursive partitioning for financial classification: the case of financial distress. The journal of finance, 40(1), 269-291.
Geng, R., Bose, I., & Chen, X. (2015). Prediction of financial distress: An empirical study of listed Chinese companies using data mining. European Journal of Operational Research, 241(1), 236-247.
Guyon, I., & Elisseeff, A. (2003). An introduction to variable and feature selection. Journal of machine learning research, 3(Mar), 1157-1182.
Holland, J. (1975). Adaptation in Natural andArtificial Systems. Ann Arbor, M]: University of Michigan Press.
JCIC. (2015). Joint Credit Information Center. Retrieved from https://www.jcic.org.tw/main_ch/index.aspx.
Jiawei, H., & Kamber, M. (2001). Data mining: concepts and techniques. San Francisco, CA, itd: Morgan Kaufmann, 5.
Jin, C., Jin, S.-W., & Qin, L.-N. (2012). Attribute selection method based on a hybrid BPNN and PSO algorithms. Applied Soft Computing, 12(8), 2147-2155.
Korol, T. (2013). Early warning models against bankruptcy risk for Central European and Latin American enterprises. Economic Modelling, 31, 22-30.
Lee, C.-C., & Huang, T.-H. (2016). Productivity changes in pre-crisis Western European banks: Does scale effect really matter? The North American Journal of Economics and Finance, 36, 29-48.
Li, H., Lee, Y.-C., Zhou, Y.-C., & Sun, J. (2011). The random subspace binary logit (RSBL) model for bankruptcy prediction. Knowledge-Based Systems, 24(8), 1380-1388.
Liang, D., Tsai, C.-F., & Wu, H.-T. (2015). The effect of feature selection on financial distress prediction. Knowledge-Based Systems, 73, 289-297.
Lin, T.-C., Yeh, C.-T., & Liu, M.-K. (2010). Application of SVM-based filter using LMS learning algorithm for image denoising. Paper presented at the International Conference on Neural Information Processing.
Liu, H., & Yu, L. (2005). Toward integrating feature selection algorithms for classification and clustering. IEEE Transactions on knowledge and data engineering, 17(4), 491-502.
Liu, J., Lin, Y., Wu, S., & Wang, C. (2017). Online Multi-label Group Feature Selection. Knowledge-Based Systems.
Maji, P., & Garai, P. (2013). On fuzzy-rough attribute selection: criteria of max-dependency, max-relevance, min-redundancy, and max-significance. Applied Soft Computing, 13(9), 3968-3980.
Majumdar, K., & Jayachandran, S. (2018). A geometric analysis of time series leading to information encoding and a new entropy measure. Journal of Computational and Applied Mathematics, 328, 469-484.
Mak, B., & Munakata, T. (2002). Rule extraction from expert heuristics: A comparative study of rough sets with neural networks and ID3. European Journal of Operational Research, 136(1), 212-229.
Mandic, D. P., & Chambers, J. (2001). Recurrent neural networks for prediction: learning algorithms, architectures and stability: John Wiley & Sons, Inc.
McMillan, E. J. (2010). Not-for-profit budgeting and financial management: John Wiley & Sons.
Michie, D., Spiegelhalter, D. J., & Taylor, C. (1994). Machine learning. Neural and Statistical Classification, 13.
Mossman, C. E., Bell, G. G., Swartz, L. M., & Turtle, H. (1998). An empirical comparison of bankruptcy models. Financial Review, 33(2), 35-54.
Ozturk, A., & Arslan, A. (2007). Classification of transcranial Doppler signals using their chaotic invariant measures. computer methods and programs in biomedicine, 86(2), 171-180.
Qi, M., Wang, T., Liu, F., Zhang, B., Wang, J., & Yi, Y. (2018). Unsupervised feature selection by regularized matrix factorization. Neurocomputing, 273, 593-610.
Quinlan, J. R. (2014). C4. 5: programs for machine learning: Elsevier.
Ruan, J., Yan, Z., Dong, B., Zheng, Q., & Qian, B. (2019). Identifying suspicious groups of affiliated-transaction-based tax evasion in big data. Information sciences, 477, 508-532.
Sahin, Y., Bulkan, S., & Duman, E. (2013). A cost-sensitive decision tree approach for fraud detection. Expert Systems with Applications, 40(15), 5916-5923.
Sammut, C., & Webb, G. I. (2011). Encyclopedia of machine learning: Springer Science & Business Media.
SFB. (2015). Securities and Futures Bureau. Retrieved from https://www.sfb.gov.tw/ch/index.jsp.
Sun, J., Li, H., Huang, Q.-H., & He, K.-Y. (2014). Predicting financial distress and corporate failure: A review from the state-of-the-art definitions, modeling, sampling, and featuring approaches. Knowledge-Based Systems, 57, 41-56.
TEJ. (2018). Taiwan Economic Journal. Retrieved from http://www.finasia.biz/ensite/
Tinoco, M. H., & Wilson, N. (2013). Financial distress and bankruptcy prediction among listed companies using accounting, market and macroeconomic variables. International Review of Financial Analysis, 30, 394-419.
Tsai, C.-F. (2014). Combining cluster analysis with classifier ensembles to predict financial distress. Information Fusion, 16, 46-58.
TWSE. (2019). Taiwan Stock Exchange. Retrieved from http://www.twse.com.tw/zh/.
Van Gestel, T., Baesens, B., Suykens, J. A., Van den Poel, D., Baestaens, D.-E., & Willekens, M. (2006). Bayesian kernel based classification for financial distress detection. European Journal of Operational Research, 172(3), 979-1003.
Vathsala, H., & Koolagudi, S. G. (2015). Closed item-set mining for prediction of indian summer monsoon rainfall a data mining model with land and ocean variables as predictors. Procedia Computer Science, 54, 271-280.
Wang, A., An, N., Chen, G., Li, L., & Alterovitz, G. (2015). Accelerating wrapper-based feature selection with K-nearest-neighbor. Knowledge-Based Systems, 83, 81-91.
Witten, I. H., & Frank, E. (2005). Data Mining: Practical machine learning tools and techniques: Morgan Kaufmann.
Witten, I. H., Frank, E., Hall, M. A., & Pal, C. J. (2016). Data Mining: Practical machine learning tools and techniques: Morgan Kaufmann.
Wu, W.-W. (2010). Beyond business failure prediction. Expert Systems with Applications, 37(3), 2371-2376.
Yahyaoui, H., & Al-Mutairi, A. (2016). A feature-based trust sequence classification algorithm. Information sciences, 328, 455-484.
Yıldırım, P. J. P. C. S. (2016). Pattern classification with imbalanced and multiclass data for the prediction of albendazole adverse event outcomes. 83, 1013-1018.
Zhou, L., Lai, K. K., & Yen, J. (2012). Empirical models based on features ranking techniques for corporate financial distress prediction. Computers & mathematics with applications, 64(8), 2484-2496.