Credit Risk Assessment Based On The Immune System Of E-commerce

The economy, science and technology have been developed rapidly, more and more people were involved in the electronic commerce, now that e-commerce transactions has the characteristics of unlimited time and space, virtual and it is also one kind of non-face to face transactions. This form of trading brings a great convenience, people need not go anywhere, trading can be made through the network at anytime. While People have the enjoyment of the convenience of e-commerce, it also accompanied by a series of e-commerce credit risk issues. Due to the asymmetric or businesses or consumers themselves, which results in a variety of problems of consumer trust in e-commerce enterprises that harm the interests of e-commerce businesses and consumers, has a bad effect on the development of e-commerce.Therefore, evaluation of credit risk of e-commerce is changing more and more important, is also currently a hot topic and focus.With the knowledge of the exploration of healthy development of e-commerce, combined with my expertise and extensive literature reference data. Especially for immune algorithm in-depth learning, I found that immune algorithm, a kind of artificial intelligence algorithms in recent years is widely developed and applied, and has achieved good efforts. Before learning together research on recommendation algorithm and differential evolution algorithm, this paper introduces the basic concepts of some of the basic concepts and biological immune system immune algorithm and artificial immune system, the achieve process of the principle and algorithm, and I found that recommended user similarity algorithm operation and differential evolution algorithm crossover could be used in this algorithm. You can find some contents of recommendation algorithms and differential evolution algorithm could be integrated into immune algorithm, that algorithm using the recommended pearson user similarity metrics to measure artificial immune network antibody-an antibody or antibody antigen fitness, the crossover and mutation of differential evolution algorithm, which could be used on the selection operating of clone process of artificial immune network antibody. Therefore, this paper take artificial immune network e-commerce credit risk on the assessment model (USAIN) a user-based similarity.By some reference, I found a set of e-commerce credit risk evaluation. Selected200cases of e-commerce corporate credit risk as the experimental data. For convenience, the data was divided into three categories,which was good, medium and poor credit. the model presented in this paper (USAIN) is set as an experimental model for comparison with the original artificial immune network algorithm (AiNet), computer simulation data, by observing the simulation image,by analyzing the accuracy of algorithm and convergence of the algorithm.I found that USAIN can completely evaluate of the credit risk of e-commerce, The success rate of100%USAIN clustering algorithm than90%AiNet algorithm, USAIN algorithm error rate is notable higher than the error rate of0.045AiNet algorithm0.14in the overall error rate. So USAIN algorithm model with strong practicality and usefulness. It Also present some new ideas to explore the assessment of e-commerce credit risk.