財務危機預測問題長久以來都是一個重要且常被廣泛討論的主題，吸引了世界各地的投資者和研究學者的關注。若想要能夠準確地進行預測，選個好的預測模型就顯得十分重要，我們從文獻中可發現，在常見的資料集下，沒有一致推薦的預測模型可供我們解決財務危機預測問題。又最近研究顯示，集成學習(ensemble learning)比單一分類器來的好且穩定，因此本研究以集成學習為基底，設計出一個新的預測模型，期望此預測模型不僅能夠對付現有的財務危機資料集，對於一個未來沒看過的資料集，也能夠準確預測。 Financial crisis problem has been important and widely studied topic. Financial crisis prediction is receiving increasing attention of stakeholders and researchers in the worldwide. If you want to predict accurately, choose a good prediction model is very important, from the literature can be found , in the common data set, there is no consistant conclusion of the prediction model for us to solve the financial crisis prediction problem, and recent studies have shown that ensemble learning is good and stable than a single classifier, thus, this study base on ensemble learning, to design a new prediction model, expect this prediction model not only deal with existing financial crisis data sets, but also predict accurately about a future data sets that we have never seen.
