Study Into Empirical Bayes On Improvement Of Credit Risk Models Accuracy

Commercial Bank in China's economic construction plays a vital role,Bank provide loans for the enterprise,console many enterprises in operation of a lack of funds, and promote the development of the our country economy overall. At present, the Bank in China's commercial credit is still the main business, and is the main way that achieved the profit by the bank. Also, that is our country enterprise to obtain an important source of business capital. The banking system has a direct influence on the national economic stability , prosperity and development. It is particularly important to study the default risk loans business. Research on the company credit risk effective method is to establish the risk evaluation model, quantitative risk. However in the model of the process, credit risk the prediction precision direct influence on the risk management level, then affect the bank risk state and profit level. Thus increasing risk metric precision of the model has the important practical significanceThis paper selects object is Chinese a-share market 2005-2009, all company annual report. It chooses financial indicators basically: operating funds/total assets, the stock market value/total liabilities, interest pre-tax profit/total assets, retained earnings/total assets, total asset turnover five indicators. Using the theory of the Logit model of financial indicators for modeling analysis, and the Bayesian theory combined experience, prior distribution the posterior distribution, likelihood function on the model to improve research, through the experience of model parameters Bayesian improve the research to credit risk measurement accuracy, has certain application value.This paper firstly introduces the credit risk metric analysis of current research status and the commonly used method, Introduces the principle of Bayesian improvement experience, then through the listed companies on Chinese financial data carried on the real diagnosis analysis, the models are presented and the improved accuracy of specific comparison, finally it discussed the real estate industry is very special in the country's economic construction recently. The paper discussed the comparison on loans default of the industry and manufacturing, then obtained the difference between default probability industry. This gives bank analysis enterprise loan defaults in dealing with different companies in industries of default risk control when a preliminary judgment..