Data Warehousing及Data Mining在營建財務及成本管理之決策支援之研究(II)

本研究以營建業之公司財務與成本因子作為研究變數，並分別以Bever(1966)之二 分類選擇法(Binary Choice)與財務危機前五年作為財務危機定義，進行模型建立；研 究方法為CART 決策樹(Decision Tree)，目的在於選取財務危機各階段預測之重要變 數以進行預測，最後再以模型穩定度與準確性評選營建業財務危機預測之模型。本研究 期能提供營建業者自我省察財務狀況之指標，並作為資金提供者之信用評估指標，進而 降低營建業財務危機之發生，鞏固營建業與金融體系財務之健全。S. Ping Ho Nation Taiwan University spingho@ntu.edu.tw Abstract The early awareness of a potential financial distress is crucial to firm’s managers for understanding their clients, suppliers and their own firms, and crucial to fund suppliers for assessing the construction firm’s credit worthiness. The purpose of this paper is to develop a dynamic prediction model for financial distress in construction industry using Data Mining. This research expects to provide construction firm managers and creditors an effective index for evaluating the credit risk a construction firm. Results show that the proposed model has higher accuracy and stability for distress prediction and can provide a more effective quantitative framework for evaluating the financial standing of a construction firm