为精确衡量卖家的整体信用和局部信用,构建多产品信用模型和单产品信用模型.在此基础上,提出大众化信用模型的基本思想,过滤不符合用户要求的卖家,提高推荐准确度.构建模型时引入评价者可信度、价格波动、惩罚因子等要素,提高信用度推荐的准确性.仿真实验验证了该模型在防范信用欺诈等方面的有效性.%This paper builds the multi-product model and single-product model to accurately measure the seller's overall credit and partial credit. On this basis, it proposes seller filtering oriented popular credit model. It can filter substandard seller and improve the recommendation accuracy. It introduces evaluator-credibility, price fluctuations, penalty factor, in order to improve the accuracy of the credit value. Simulation results verify the model's effectiveness to prevent credit fraud.
展开▼
