【摘要】 针对含有大量偏好信息的不完备信息系统,提出一种基于β-先验概率优势关系的粗集模型。首先给出β-先验概率优势关系模型的定义,其次讨论了新模型的一些相关性质及其近似分类质量。β-先验概率优势关系模型具有一定的容错能力,通过设置参数值来提高对数据预测、分类的合理性,从而达到与主观认知相一致的分类效果。实例分析进一步验证了新模型的优越性。更多还原

【Abstract】 A rough set model based on β-prior probability dominance relation is proposed to solve some problems in the incomplete information system with lots of preferential information. Firstly, the definition of β-prior probability dominance relation is proposed. Secondly, some related properties and approximation classification quality about the new model are discussed. The new model can improve accuracy of data forecast and classification by setting parameters’ values. The classification achieved by the new model is almost as same as the classification judged by subjective cognition. The new model softens the dominance relation in some extent and increases the robustness in data analyzing and disposing. At last, the new model proves its superiority by analyzing an example.更多还原