Studies on the Significance of System Attributes with the Information Uncertainty Methods
基于不确定性问题研究方法评价系统参数的重要性

Selecting useful attributes in pattern recognition is studied with the statistical method on the principle of entropy reducing information uncerainty and the approximation classification method on the concepts of rough sets in the paper. A comparison of rough sets and statistical methods in evaluating the significance of system attributes is made. Finally, the uniformity of rough sets and statistical methods in information uncertainty is discussed on selecting system attributes.