A Credit Rating Model For Individual Based BP Network And Genetic Algorithms

In order to improve the risk management of the financial institutions, the Bank for International Settlement announced the New Basel Capital Accord (Basel II), which will become effective at the end of 2006. The Basel II are concerned with the minimum capital requirements for losses resulting from credit, market and operational risk with implications on a risk management framework within financial institutions. To date, the credit risk serve as major risk of most banks'risk causes. Once the credit risk occurs, the bank will suffer a huge sum loss. The system has the capability of providing correct sign for warning the bank to avoid or mitigate, when debtor's probability of default is higher then threshold.The academic circles put forward various classification models used for the risk appraisal, these models can be classified into two major types: the traditional statistics model and artificial intelligence model. The traditional statistics model mainly according to the multiple statistics analysis. The basic way of thinking is , according to some samples in every category(violation type, normal type) which have already been mastered in history, summarize a classification regulation from it, and establish a discretion formula to used for the classification of the new samples. According to the form of discretion function and the different supposition of sample's distribution, the main models are: multiple regression analysis model, multiple discretion analysis model(MDA), the logit analysis model, nearest neighbor classification method. Among them , MDA and the logit analysis model are in most widely application. Most obvious superiority is that it have obvious explanation, and disadvantage is that prerequisite is too strict. With the development of information technology, the...