多尺度随机信息系统的粗糙近似关系

Rough approximation relations of multi-scalerandom information systems

传统随机信息系统中的条件属性都是单一尺度,而现实生活中一个对象却有多方面的刻画.为更好地分析数据,在该系统里结合多尺度信息系统中的粒度转换函数,给出多尺度随机信息系统的概念,并研究不同尺度下的信息粒和粗糙上下近似算子之间的关系及最优尺度选择.

In the classical random information systems,the condition attribute has a single scale,but in the real life,an object has many aspects of description.In order to analyze the data better,combining the granularity transformation function in the multi-scale information systems,we give the concept of the multi-scale random information systems,then study the relationship of information grains and rough upper lower approximation op⁃erators between different scales of the systems and the optimal scale selection.