Optimization Of Personal Credit Evaluation Method And Its Application In China

Personal credit evaluation is an important component of risk management, with the development of society, personal credit evaluation has been obtained widely attention and application. It can be transformed into two-class classification problems, there are many methods commonly used for two-class classification problem, such as regression algorithm, nearest neighbor algorithm, genetic algorithm, neural network, support vector machine and so on. And the support vector machine has become a hotspot in the research of personal credit evaluation. Based on the support vector machine, we get the support matrix machine to deal with the matrix data classification. In order to solve the nonlinear classification problem, this paper also cited the kernel function of the matrix data, and then combined the matrix data kernel function and support matrix machine, it gets a new method to deal with the personal credit evaluation.The contents and results in this paper as follows:A brief description of the classification problem and the theory of support vector machine.The personal credit evaluation data is the traditional vector data, and the vector data has two shortcomings:one is lost the relationship information between the index; two is not in a dynamic, these problems will greatly affect the final credit assessment results.It aims at the shortage of the vector data, we propose a matrix type data of the individual credit assessment, and put forward two kinds of methods to construct the matrix data, of course, the two methods are not perfect, we analyzed the two methods and compared the advantages and disadvantages of them and credit evaluation problem.The kernel support matrix method proposed by us, and tested in the foreign credit evaluation data set German, Australian and the domestic housing loan credit data set, and the results prove the effectiveness of our new approach...