An Improved K-modes Clustering Algorithm Based On Rough Sets

K-modes clustering algorithms are widely used in artificial intelligence,data mining and other fields.The traditional K-modes clustering algorithm has good clustering effects,but it has many problems,such as multiple iteration times,a large amount of computation,and is easy to be interfered by redundant attributes.Besides,it only uses a simple 0-1 matching method to define the distance between two sample attribute values,without fully considering of the influence of each attribute on the clustering results.To solve these problems,this paper introduces classical Rough entropy into K-modes algorithm.Firstly,the rough set attribute reduction algorithm is used to eliminate redundant attributes and determine the importance of each attribute.Then,the weight of each attribute is determined by using classical rough entropy,and defines a new in-class distance to optimize the objective function of K-modes clustering algorithm.The proposed algorithm is compared with the traditional K-modes clustering algorithm on four public data sets.And the experimental results show that the proposed algorithm has higher clustering accuracy than the traditional K-modes clustering algorithm,but the difference in clustering purity is small,and the running time is relatively longer.In addition,in order to further verify the applicability and practicability of the proposed algorithm,the proposed algorithm is applied to image segmentation.Image segmentation is an important step in image processing,which divides an image into many different parts for more detailed processing.It can be used to extract objects in images,detect edges,extract features and so on.The experimental results on SRCNN data set show that the number of clustering can be accurately determined by using the peak number of gray value,which has high flexibility and can improve algorithm efficiency.In this paper,the improved K-modes image segmentation algorithm based on rough sets can segment images well.