Improving on a rapid attribute reduction algorithm based on neighborhood rough sets

The neighborhood rough sets, which can deal with continuous attribute values directly without data discretization, is easy to understand. So it is widely used in continuous attributes reduction. However, existing methods need to spend a lot of time to process large samples data and thus more effective method needs to be proposed. In this paper, several mathematical properties of neighborhood rough sets are analyzed. The algorithm FARNeMF (Forward Attribute Reduction Based on Neighborhood Rough Sets and Fast Search) in the literature [1] will be improved. By this new algorithm, the comparison times of samples in computing positive regions and neighborhoods is reduced. Finally, experimental results show that the proposed method is more effective than existing methods.