粗糙集理论应用中的离散化方法综述

粗糙集理论是一种有效处理不确定、不精确、不完备信息的数学工具。但是传统的粗糙集理论只能对数据库中的离散属性进行处理，而绝大多数现实的数据库既包含了离散属性，又包含了连续属性。针对粗糙集理论的这些缺陷，综述了几种比较常用的离散化算法，并在此基础上，引申出一些启发式离散算法。最后指出，现有的离散化算法都会或多或少地损失部分信息，目前还没有确定的评判准则评论哪一种方法更好、更合适；离散化方法的一个有效思想就是“最低限度地减少信息丢失、保持数据分类能力和使最终得到的离散化最小（分类器最小）”的原则。

Overview of the Discretization Methods in the Application of Rough Set Theory

The rough set theory is an implement of math to deal with the information of indefiniteness,impreciseness and incompleteness effectively.The traditional rough set theory can only deal with the discrete features of the datasets,but the most of the realistic datasets include not only the discrete features but also the continuous features.To the disadvantages of the rough set theory,several popular discretization methods are overviewed,and based on this,some heuristic methods provided.At last,it is pointed out that all of the discretization method can lose part of information more or less,but at present there is no certain method to determine which method is better,or more fit.The effective idea of discretization method is the principle of "minimizing the loss of the information at the most extent,keeping the ability of data classification and making the discretization minimally(the classifier minimize)".