Randomly sampling based fuzzy rough reduction

Now most methods of the existing fuzzy rough based attribute reduction can´t be applied to large scale data sets because of high time and space consuming. To overcome this problem, we proposed a randomly sampling based method to decrease the computing consumption. First, we propose some pseudo concepts of fuzzy rough sets, such as the pseudo lower approximation and pseudo attribute reduction, which is not exactly, but approximately equal to classical concepts, and then it drastically accelerate computation and significantly reduce space consumption. The numerical experiments show that the random sampling algorithm is significantly efficient both in time and space with little loss of information.