Summary: | 碩士 === 輔仁大學 === 資訊管理學系 === 97 === Many countries in the world have developed the prediction models in order to face the effects of the globe financial crisis. In Taiwan, the architecture of CAMELS has been used to build the system of the prediction. Following the financial liberalization, Taiwan government releases gradually the restriction of the banking policy which is expected to enhance the performance of the banks and make banks more prosperous in the market. However, the measure does not increase dramatically the banks’ competition and the problems of those financial intuitions occur continually on the contrary. In addition, with the loose financial policy, the banks confront the risk that encourages the developing countries to establish the risk-oriented prediction methods which have become common trends worldwide. This study built a prediction model by taking financial data of 49 Taiwanese local banks for the research samples, including the financial reports quarterly between 2000 and 2008. That system of prediction follows the principle of CAMELS and selects the factors which have been highlighted in the past researches. There are four techniques of data mining that are implemented for this study.
The goal of this study is to predict the probability of operation risk in Taiwanese local banks and apply decision tree, Bayesian classification, Logistic regression and back propagation neural network to analyze the specific public data. The results of prediction have been evaluated precisely and expect to provide the reliable references for the banking intuitions in supervising issues and improving the performance.
|