在過去數十年間，人們面臨著動盪的金融體系，承受著企業違約的風險，在面對金融危機時遭受巨大的損失。雖然有第三方公平機構對於市面上的公司有著對應的企業信用風險評等，然而卻缺乏即時性，需要等到財務報導公布後一段時間內才會發布其信用風險評等，無法讓財務報導使用者獲得第一手的消息並進行決策分析。
　　基於上述提到的問題，本研究針對全台灣上市櫃公司進行研究，採用隨機森林模型分析財務報表中的各項變數，建立對於台灣上市櫃公司信用風險的預測模型，研究結果顯示，透過財務報表內與償債能力以及獲利能力的相關財務數字，可以有效預測公司的信用風險，並且提供使用者即時的預測風險結果。

　　In the past decades, people faced unstable financial system, and suffered damage or loss by the company which didn’t fulfill contract obligations. Although we had a fair institution to evaluate which company had a high credit risk, but the results only showed up after a few months of financial statement release. Financial statement users can’t make decisions by the information in time.
　　Based on the problems we mention, this paper conducts a study on the listed company in the Taiwan, analysis the number in the financial statement, and use random forest model to predict credit risk. We find that according to the ability of debt-paying and profit earning, can predict credit risk accurately, and provide the real time result.

誌謝 i
摘要 ii
Abstract iii
壹、緒論 1
一、研究背景與動機 1
二、研究目的 3
三、研究方法 4
貳、文獻探討 5
一、信用風險 5
二、國際信用風險評等機構 7
三、台灣企業之信用風險評等 8
四、機器學習 11
五、決策樹 12
六、隨機森林 13
七、自適應增強(AdaBoost) 13
八、K-近鄰演算法 14
十、預測信用風險 15
十一、模型評估指標 18
十二、財務報表變數構面 20
參、研究方法 23
一、研究架構 23
二、移動視窗(Sliding Window) 25
三、財務報表變數介紹 25
四、資料降維 26
五、實驗環境設定 27
六、實驗設計 27
肆、實驗結果 31
一、 第一次實驗 31
(一)第一次實驗(無降維) 31
(二)第一次實驗(一次降維) 32
(三)第一次實驗(二次降維) 36
(四)第一次實驗(三次降維) 39
(五)變數分析 41
(六)結論 45
二、 第二次實驗 45
(一)第二次實驗(無降維) 46
(二)第二次實驗(一次降維) 48
(三)第二次實驗(二次降維) 51
(四)第二次實驗(三次降維) 54
(五)變數分析 55
(六)結論 60
伍、結論與討論 61
一、結論與研究貢獻 61
二、研究限制與未來展望 62
參考文獻 64

Ahmadyan, A. 2018. "Measuring Credit Risk Management and Its Impact on Bank
Performance in Iran," Marketing and Branding Research (5), pp. 168-183.
Anshul Kumar Singh, S. K. M. 2013. "Ordered Response Comparison between Fitch,
Standard & Poor, Moody Ratings," in: Department of Humanities and Social
science. Kanpur ,India: Indian Institute of Technology.
Antonio M. Mora García, P. A. C. V., Juan J. Merelo Guervós, Eva Alfaro Cid, Anna
I. Esparcia-Alcázar, Ken Sharman. 2008. "Discovering Causes of Financial
Distress by Combining Evolutionary Algorithms and Artificial Neural
Networks," in: Proceedings of the 10th annual conference on Genetic and
evolutionary computation, M. Keijzer (ed.). Atlanta GA USA.
Bissoondoyal-Bheenick, E. 2004. "Rating Timing Differences between the Two
Leading Agencies: Standard and Poor’s and Moody’s," Emerging Markets
Review).
Branston, L. 1995. "Mexican Crisis Underlies Agency Pitfalls," Financial Times).
Breiman, L. 2001. "Random Forests," Machine Learning (45), pp. 5-32.
Claude B. Erb, C. R. H., Tadas E. Viskanta. 1995. "Country Risk and Global Equity
Selection," Journal of Portfolio Management), pp. 74-83.
Constantin Mellios, E. P.-B. 2006. "Which Factors Determine Sovereign Credit
Ratings?," The European Journal of Finance (12).
Cressey, D. R. 1953. "A Study of the Social Psychology of Embezzlement," Other
people's money).
Danilo Bzdok, N. A., Martin Krzywinski. 2018. "Statistics Versus Machine Learning,"
Nature Methods (15), pp. 233-234.
Dittrich, F. 2007. "The Credit Rating Industry: Competition and Regulation," in:
Department of Economics. University of Cologne.
Flavio Barboza, H. K., Edward Altman. 2017. "Machine Learning Models and
Bankruptcy Prediction," Expert Systems with Applications (83).
Frank Packer, N. A. T. 2012. "Rating Methodologies for Banks," BIS Quarterly
Review).
Gongde GuoHui, W. B., BiKieran Greer. 2003. "Knn Model-Based Approach in
Classification," in: OTM Confederated International Conferences "On the
Move to Meaningful Internet Systems". Berlin.
Harjani Rezkya Putri, A. D. 2019. "Prediction of Financial Distress: Analyzing the
Industry Performance in Stock Exchange Market Using Data Mining," in: 2019
16th International Conference on Service Systems and Service Management
(ICSSSM). Shenzhen, China: IEEE.
63
Issam El Naqa, a. J. M. 2015. What Is Machine Learning?
J A Hanley, B. J. M. 1982. "The Meaning and Use of the Area under a Receiver
Operating Characteristic (Roc) Curve.," Radiology (143).
J. Galindo, P. T. 2000. "Credit Risk Assessment Using Statistical and Machine
Learning: Basic Methodology and Risk Modeling Applications,"
Computational Economics (15), pp. 107-143.
James C. Van Horne, J. M. W. 2005. "Prinsip-Prinsip Manajemen Keuangan," Edisi
Keduabelas. Jakarta: Salemba empat).
Leo Breiman, J. H. F., Richard A. Olshen, Charles J. Stone. 1993. "Classification and
Regression Trees," Chapman and Hall).
M. I. Jordan, T. M. M. 2015. "Machine Learning: Trends, Perspectives, and
Prospects," Science).
Mahdi Moradi, M. S., Mohammad Ebrahim Ghorgani, Hadi Sadoghi Yazdi. 2013.
"Financial Distress Prediction of Iranian Companies Using Data Mining
Techniques," Organizacija (46).
Mihaelajméno, M. O. 2015. "Credit Rating Agency Performance in Terms of Profit,"
Procedia Economics and Finance (30).
Quinlan, J. R. 1985. "Induction of Decision Trees " Machine learning).
Quinlan, J. R. 2014. C4. 5: Programs for Machine Learning. Elsevier.
Ruibin Geng, I. B., Xi Chen. 2015. "Prediction of Financial Distress: An Empirical
Study of Listed Chinese Companies Using Data Mining," European Journal of
Operational Research (241).
S.R. Safavian, D. L. 1991. "A Survey of Decision Tree Classifier Methodology," IEEE
Transactions on Systems, Man, and Cybernetics), pp. 660-674.
Schapire, R. E. 2013. Explaining Adaboost. Springer.
Tannu Chauhan, S. R., Samrath Malik, Pushpa Singh. 2021. "Supervised and
Unsupervised Machine Learning Based Review on Diabetes Care," in: 2021
7th International Conference on Advanced Computing and Communication
Systems (ICACCS). Coimbatore, India.
Tomy Rizky Izzalqurny, B. S., Abdul Ghofar. 2019. "Relationship between Financial
Ratio and Financial Statement Fraud Risk Moderated by Auditor Quality,"
International Journal of Research in Business and Social Science (8), pp. 34-
43.
Victoria Y Fan, D. T. J. a. L. H. S. 2018. "Pandemic Risk: How Large Are the
Expected Losses?," Bull World Health Organ), pp. 129-134.
W.Goodell, J. 2020. "Covid-19 and Finance: Agendas for Future Research," Finance
Research Letters (35).
Xiaobing Huang, X. L., Yuanqian Ren. 2018. "Enterprise Credit Risk Evaluation
64
Based on Neural Network Algorithm," Cognitive Systems Research).
Yoav Freund, R. E. S. 1996. "Experiments with a New Boosting Algorithm," in:
International Conference on Machine Learning.
You Zhu, L. Z., Chi Xiea, Gang-Jin Wang, Truong V.Nguyen. 2019. "Forecasting
Smes' Credit Risk in Supply Chain Finance with an Enhanced Hybrid
Ensemble Machine Learning Approach," International Journal of Production
Economics (211), pp. 22-33.
Yu-Pei Huang, M.-F. Y. 2019. "A New Perspective of Performance Comparison
among Machine Learning Algorithms for Financial Distress Prediction,"
Applied Soft Computing (83).
Yuelin Wang, Y. Z., Yan Lu, Xinran Yu. 2020. "A Comparative Assessment of Credit
Risk Model Based on Machine Learning ——a Case Study of Bank Loan
Data," Procedia Computer Science (174), pp. 141-149.
方豪、李彥賢、姜廷宜. 2017. "廠商融資對銀行信用風險之影響," 兩岸金融季刊
(5).
王育慧. 2007. "最高法院九十一年度台上字第三○三七號判決及臺灣高等法院八
十八年度上重訴字第三九號判決之評釋（東隆五金案）－以連續交易操
縱行為與內線交易為範圍," 臺北大學法學論叢 (61), pp. 43-78.
古永嘉、李在僑、羅玉惠. 2009. "加入公司治理指標建構信用評等預測模型之研
究," 當代會計 (10), pp. 131-162.
何思湘. 2002. "美國安隆公司倒閉事件對我國相關立法之啟示（一）," 會計研究
月刊 (199), pp. 17-28.
吳珮甄. 2013. "台灣上市櫃食品公司之信用風險預測," in: 風險管理與保險研究
所. 國立高雄第一科技大學, p. 59.
吳清在. 2022. "產業別財務危機預警模型建構之研究," in: 國家發展研究所. 臺北
市: 國立臺灣大學.
呂俊杰. 2019. "企業信用風險預警模型探討—以門檻邏吉斯迴歸模型分析," in:
國際財務金融碩士在職專班. 國立臺北大學, p. 62.
呂建錡. 2021. "以 ai 建立煤炭性質預測、生質能氣化、與生質能發電之模式—決
策樹回歸、Svr、Ann 之比較," in: 機械工程學系所. 國立中興大學.
李欣海. 2013. "隨機森林模型在分類與迴歸分析中的應用," 應用昆蟲學報 (4),
pp. 1190-1197.
邱文昌. 1999. "我國建立信用評等制度之規劃與檢討," 證交資料).
紀淑芬. 2021. "企業財務危機預警影響因素之研究 - 以台灣電子業上市櫃公司
為例," in: 會計資訊系碩士班. 臺中市: 嶺東科技大學.
張媚雅. 2009. "公司現金持有與現金持有價值-追溯亞洲金融風暴," in: 財務金融
所. 國立中正大學.
許鳳芸. 2010. "利用羅吉斯迴歸建立台灣上市櫃公司財務預警模型," in: 財務金
65
融研究所. 臺南市: 國立成功大學.
郭佩禎. 2019. "高科技業公司治理與信用風險模型之實證研究," in: 國際企業管
理學系碩士在職專班. 中國文化大學, p. 71.
郭蔚楓. 2018. "邏輯斯迴歸與 roc 曲線整合應用於探討道路鋪面破壞影響因素—
以台 26 線為例," in: 土木工程學系. 屏東縣: 國立屏東科技大學.
陳伯松. 2007. "看力霸弊案中，財務報表如何唬弄！ 財報還原真相 營私舞弊
難逃," 會計研究月刊 (258), pp. 75-81.
歐再添. 2003. "企業財務危機預測--以產業別建構 logistic 預警模型," in: 企業管
理系. 臺北市: 國立臺灣科技大學.
賴振東. 2012. "以公司治理因素分析上市企業信用風險," in: 經營管理研究所在
職專班. 國立虎尾科技大學, p. 107.
嚴奇峰；林雅惠；楊燿鴻. 2012. "博達公司個案－從利益收斂假說的觀點探討,"
產業管理評論 (5), pp. 55-72.
蘇瑩珊. 2012. "全球金融風暴下美國促進就業政策之探討 -兼與「新政」時期政
策比較
" in: 勞工關係學系暨研究所. 國立中正大學.