Abstract: Variable precision approach has a very good noise immunity ability in the data treatment，the random set mapping can naturally lead to a neighborhood relation with proper property，and their combination can satisfy the great requirement of an extensive application background.In this paper，the variable precision rough set model based on random sets of real-values information system is studied，the approximation operators of variable precision probabilistic rough set based on the random sets are constructed and some related properties of this model are discussed.In the mean time，the composition problem of this rough set model and its some properties are also researched.The results of this paper are very helpful to the researches of architecture methods of rough set model of some complex information systems.

Key words: variable precision rough set, random set, probabilistic measurement, composition model

摘要： 变精度方法对噪声数据具有良好的抗干扰能力，随机集映射本质上给出了一个具有良好性质的邻域关系，两者相结合能应对更广泛背景的应用需求。从随机集出发研究实值信息系统的变精度粗糙集模型问题，定义了以概率测度为基础的变精度粗糙集近似算子，并讨论这种模型的相关性质。同时也讨论了实值信息系统的变精度粗糙集模型的合成问题。结果有助于研究复杂系统的知识和信息数据处理的粗糙集模型的构造问题。

关键词: 变精度粗糙集, 随机集, 概率测度, 合成模型