부산대학교 도서관

In order to adapt to the actual situation of the complex network environment, and make the credit rating results be more dynamic and authentic. We divide the sellers' transaction history into several time frames, and define four concepts including recent credit, long-term credit, cumulative negative credit and evaluator credibility. The model is more suitable for the reliability and authenticity of the credit compared with the traditional credit rating model. Through studying the existing credit rating platform, this paper analyzes the existing problems and the advantages of the system itself, and establishes a new credit model by taking many kinds of factors into consideration, which include various objective factors, such as the amount and completion time of the transaction, the buyer's credit degree, etc. In the end, the implementation of the credit rating model is given based on the above factors. [ABSTRACT FROM AUTHOR]