Fast Reduction Algorithm Research Based on k-Nearest Neighbor Fuzzy Rough Set

Now a lot of generalized models of rough set are proposed by introducing some parameters to deal with the problems with noise. Traditional reduction algorithms are designed to find the minimum subset which keeps the information invariant. However, there is an obvious weakness that the algorithms have to be executed from the beginning on different parameters. This paper introduces the theoretical results of nested structure into the robust fuzzy rough set(i.e., k-nearest neighbor fuzzy rough sets), and then designs a fast reduction algorithm based on given reduction by using the nested structure. The main contribution of the proposed algorithm is that it can quickly find a reduction on different parameters when one reduction on certain parameter is already given. The numerical experiments verify that the executing time can be significantly saved through using fast reduction algorithm and demonstrate that the proposed algorithm is feasible and effective.