基于Dirichlet分布的电子商务信誉评估模型

Reputation Rating Model Based on Dirichlet Distribute in E-Commerce

  • 摘要: 在已有的信誉理论基础上,结合Dirichlet概率分布以及“差评敏感”的评估原则,提出了一个适用于电子商务的信誉评估模型。该模型基于Dirichlet概率密度函数,从概率的角度客观地反映了实体集的评估结果,根据评估实体输入的离散评估值,得出连续的信誉值。同时,“差评敏感”处理保证了模型对差评的敏感性,提高了模型的谨慎度,而且随着评估次数的增加,模型对评估实体具有自适应性。模拟实验表明,该模型有着良好的评估特性,适用于电子商务信誉评估的环境。

     

    Abstract: Based on the research of the reputation theory, a reputation rating model based on Dirichlet probability density function and “ad-sensitive” rating rule is proposed in e-commerce. The model takes discrete ratings of entity as input, and derives the continuous measures. “Bad-sensitive” rating rule makes the model sensitive and more cautious to the bad ratings and. The model is self-adaptive as the increasing of ratings. Experiments show that the model has several good characters for the reputation rating in e-commerce.

     

/

返回文章
返回