財務危機預警模型係利用企業財務與非財務資訊以觀察企業整體之營運狀況，藉此判別企業是否有財務危機之疑慮。集團企業的組織結構與經營方式的獨特性會導致財務報表與績效與非集團企業有所差異，因此，本研究擬將集團企業的獨特性納入DEA-DA模型，期望以DEA-DA最小化錯誤分類個數的特性來建構出高準確度的預警模型，並針對不同的景氣情況加以分析探討。實證結果發現，在不區分景氣下，前一年至前三年預警模型的估計準確度為86.7%、84.3%與79.5%。透過景氣的區分，的確能使預警模型的估計準確度在危機前一年由86.7%提昇至92%左右。在景氣下降期間，集團公司除了平時要注意應收帳款的回收能力，而在愈接近危機時點時，償債能力的維持是最重要的。但在景氣上升期間，則需注意集團公司是否隨景氣好轉過度盲目增加銷貨或過度投資，反而忽略獲利能力與營運效率，無法為股東創造真正的價值。

Financial distress prediction model is generally using financial and non-financial information to observe the whole operation condition of a company, and discriminate whether a company is financial distress entity or not. The uniqueness of group organization and group business operation contribute to the significant difference with non-group business in the financial statements and performance. Therefore, this study include the uniqueness of business group into the DEA-DA model and utilize the characteristic of the DEA-DA model-the minimization of misclassifications, to construct a more accurate financial distress prediction model. Furthermore, this study also considers the influence of different economic climates on the financial distress prediction model. The empirical results show that the degree of year by year accuracy on the financial distress prediction model is 86.7%, 84.3% and 79.5% under all economic climates. We also find that by discriminating the business cycle, the accuracy of the financial distress prediction model increase from 86.7% to 92% one year prior to the failure. In the economic recession period, group-affiliated firms should pay more attention to the turnover of accounts receivable and solvency for avoiding the occurrence of financial distress, especially closing to the distress point of time; in the economic expansion period, group-affiliated firms should focus on the profitability and the improvement of the operation efficiency to create real value for the shareholders, not to boost up too many sales and overly investing with the better economic climate.