應用資料探勘方法於企業財務經營危機診斷模式建立

Using Data Mining Method to Construct a Financial Crisis Model

張念慈(Nian-Tsz Chang)

指導教授 ： 林榮禾

國立臺北科技大學/管理學院/商業自動化與管理研究所/碩士(2008年)

摘要

近年來由於企業面臨激烈競爭與多樣且多變的經營環境，管理一旦成效不彰則易引發財務危機甚至於公司倒閉的情況，進一步造成社會的不安定與金融秩序的嚴重衝擊。因此，過去幾年來，有許多學者們紛紛投入建構危機預警模型相關研究，以達成企業經營成效之提升與降低企業營運成本降低之目的，本研究欲建立一套企業財務危機診斷模式，期望能提供監督機構、投資人、企業營運單位，提早在危機發生前有所警覺。
企業財務危機診斷模型是利用資料探勘方法建構，其分為四部分：第一部分為建立單一診斷模型和危機再確認模型。第二部分則是建立多重組合危機預測模式以判別分類結果是否一致，若結果不一致，則分別利用案例式推理模式 (Case-base Reasoning) 和投票機制預測模式做診斷。第三部分，是建立簡化型單一危機預測模式和多重組合危機預測模式，最後一個部分利用基因演算法 (Genetic Algorithm, GA) 在眾多方案中找出最佳化組合方式。
研究結果顯示透過危機再確認模式和不一致診斷模式可以得到較佳的預測效果，再藉由最佳化模式可以較快找到一套最適財務經營危機預測模式。這模式能提供企業在危機中如何變革求生的參考與建議，進而提升企業經營之競爭力。

關鍵字

危機診斷系統 ； 預警模式 ； 資料探勘 ； 統計分析

並列摘要

In recent years, enterprises worldwide are confronted with fierce competition and a constantly changing and increasingly difficult business environment. In the face of such challenges, ineffective management can easily cause financial distress or even bankruptcy that in turn can trigger social instability and impacts the financial order. Therefore, in recent years many scholars have invested considerable efforts on the establishment of financial crisis prediction model to help enterprises enhance management effectiveness and reduce operating costs. This study aims to construct a financial crisis model to predict financial distress and to provide enterprises, investors, and the Government with an effective tool in financial monitoring and business management.
A financial crisis model based on data mining method can be divided into four parts. The first part utilizes a single diagnosis model and a crisis reconfirmation model. The second part develops a multiple-classifier financial crisis prediction model to examine the consistency of the classification results. If any inconsistency occurs, either the case-based reasoning system or the majority voting approach is adopted to perform the prediction. The third part builds a simplified single model and integrates the multiple classifiers of financial crisis prediction. The last aspect utilizes GA to find optimal combinations.
Research results indicate that the crisis reconfirmation and inconsistency-based models obtain better prediction results. The optimization model can then be applied to speed up the identification of an optimal financial crisis model. The proposed model can be used to provide suggestions for enterprises to overcome crises and to enhance their competitiveness.

並列關鍵字

Crisis diagnosis system ； Prediction ； Data mining ； Statistics

參考文獻

[18]黃振豐、呂紹強，「企業財務危機預警模式之研究－以財務及非財務因素建構」，當代會計期刊，第一卷，第一期，2000，第19-40頁。

[14]陳柏榮，以限制規劃程式構建投資組合決策支援系統之研究，未出版碩士論文，國立交通大學運輸科技與管理學研究所，2002。

[28]Altman, E.I., “Financial ratios, discriminant analysis and the prediction of corporate bankruptcy,” Journal of Finance, vol. 22, 1968, pp. 589-609.

[29]Altman, E. I., Predicting financial distress of companies: revising the Z-score and ZETAR models. Adapted and updated from Altman (1968) and Altman (1977), 2000.

[30]Beaver, W.H., “Financial ratios as predictors of failure,” Journal of Accounting Research, 1966, pp. 71-111.