企業財務危機預警模型的研究一直都是政府機關、金融業者、企業單位及投資者所關注的議題，而過去的研究大多以財務比率來建構模型，其財務比率多由財務報表項目以及業務統計之項目所計算得到，因此本研究除了將財務比率列為解釋變數，同時也包含財務報表之會計項目以及業務統計項目。
本研究以財務報表之會計項目、業務統計項目以及財務比率等財務資訊建構台灣壽險公司破產預測模型，資料期間為2005年至2014年，共有292個樣本，其資料來源為財團法人保險事業發展中心之公開資訊，模型則使用三種大數據資料探勘之分類演算法，分別為羅吉斯迴歸(Logistic regression)模型、決策樹(Decision Tree)模型以及支持向量機(SVM)模型，以 10-fold 交叉驗證法避免模型過適(Overfitting)，並分別比較此三種模型之效度(Validation)，以及找出影響公司破產之主要變數。
根據本研究實證發現，不採用公司破產後資料之支持向量機模型較可信且較符合實際狀況，並考慮正常公司與破產公司之資料不平衡問題，「資本公積」、「各項責任準備金對資產比率」、「保費收入-團體險-傷害保險」以及資產負債表之「其他應收款」與「現金及銀行存款」，此五個變數最為顯著。

Corporate Financial Crisis Early Warning Model has been an issue that brought up attention from government agencies, financial operators, enterprises and investors. Previous studies generally used add financial ratios as independent variables to predict corporate financial distress. However, financial ratios are all computed by the subjects of financial statements and the business statistics items. This study not only includes subjects of financial statements, but also business statistics items as explanatory variables to construct the model.
In this study, the subjects of financial statements, business statistics items and financial ratios are used as explanatory variables to construct the model of Bankruptcy Prediction for Taiwan Life Insurance Companies. The samples of this research are selected from Taiwan Insurance Institute databases from 2005 to 2014 and a total of 292 samples. This study uses three classifications of data mining methods respectively, including Logistic regression model, decision tree model and support vector machine model and uses 10-fold cross validation to avoid overfitting the model. Lastly, we will compare the accuracy and specificity of the three classification models through the research.
The result of empirical analysis shows that SVM model without samples of bankrupt company is the more reliable method for predicting the probability of bankruptcy for Taiwan Life Insurance Companies. Considering that the number of normal companies and insolvent companies in the samples is unbalance, "Capital Surplus", "the ratio of reserves and assets ", "Premium – Group Injury Insurance ", and the balance sheet of "Other Receivables" and" Cash and Cash in bank" are significant explanatory variables.

1. 張士傑，(2015)。台灣保險市場發展、監理與評論。出版地：財團法人台灣金融研訓院
2. Foster Provost, T. F. (2013). Data Science for Business: What You Need to Know about Data Mining and Data-Analytic Thinking. Published by O’Reilly Media,
Inc., 1005 Gravenstein Highway North, Sebastopol, CA 95472.
3. Beaver, W. H. (1966). Financial ratios as predictors of failure. Journal of accounting research, 4, 41. doi:10.2307/2490171
4. Altman, E. I. (1968). "Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy." the journal of Finance 23(4): 21.
5. Beaver, W. H. (1966). "Financial ratios as predictors of failure." Journal of accounting research 4: 41
6. Altman, E. I. (1968). "Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy." the journal of Finance 23(4): 21.
7. Beaver, W. H. (1966). "Financial ratios as predictors of failure." Journal of accounting research 4: 41.
8. Foster Provost, T. F. (2013). Data Science for Business: What You Need to Know about Data Mining and Data-Analytic Thinking.
9. Martin, D. (1977). "Early Warning of Bank Failure: A Logit Regression Approach." Journal of Banking and Finance: 76.
10. Odom, M. D., Sharda, R. (1990). A neural network model for bankruptcy prediction. 1990 IJCNN International Joint Conference on Neural Networks. 2: 6.
11. Ohlson, J. A. (1980). "Financial Ratios and the Probabilistic Prediction of
Bankruptcy." Journal of Accounting Research 18(1): 23.