An Improved Uncertainty Measure Method of Rough Sets

The rough entropy always reduce with the subdivision of knowledge particles when it is used to measure the uncertainty of rough sets, no matter what the knowledge particles come from positive domain, negative domain or boundary domain. so the method rough entropy can not reflect the relationship between uncertainty and knowledge particles well. To solve this problem, an improved method based on rough entropy is proposed. This method is proved that it satisfies the basic standards and extended standards of uncertainty measure of rough sets. Then come to a conclusion that the uncertainty of rough sets will reduce only when knowledge particles which belong to boundary domain are subdivided. And this improved uncertainty method is applied into the fuzzy-rough algorithm based on fuzzy-rough theory.