Abstract

The classic Pawlak rough set model is only suitable for processing discrete data, not continuous data directly. Pawlak rough set model needs discretization when processing continuous data, which will cause the loss of information in the data set. In this context, Hu proposed the neighborhood rough set (NRS) model, which greatly expanded the application range of rough set theory. Attribute reduction is one of the main applications of the neighborhood rough set theory, but the existing attribute reduction algorithms based on neighborhood rough set have high computational cost and long running time. To address this problem, a fast attribute reduction based on $k$ -means granular ball neighborhood rough set ( $k$ -GBNRS) is proposed by introducing granular ball computing into neighborhood rough set and combining with an efficient clustering method. The proposed $k$ -GBNRS method not only runs fast, but also generates an adaptive radius, which fits the data set well. Experimental results show that this method is more efficient than the latest NRS method.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call