粗糙集作为数据挖掘工具,主要通过分类数据得到预测型知识,但分类规则过于严格,使得挖掘结果可能会损失一些有价值的规则,本文引入带不确定性因子的决策系统UFDS,在该系统中根据统计结果和领域知识为每一对象赋以不确定度k和重要度p,并对传统等价类划分进行扩充,成为重要类和负类,在此基础上提出了带不确定因子的属性约减算法。 Rough set as a data mining tool, mainly through the classification of data to obtain predictive knowledge, but the classification rules are too strict, making the mining results may lose some valuable rules. In this paper, the UFDS with uncertainty factor decision-making system, in this system Based on the statistical results and domain knowledge, each object is assigned an uncertainty k and an importance p, and the traditional equivalence class partition is expanded to become an important class and a negative class. Based on this, an attribute with an uncertain factor is proposed Reduction algorithm.