Research of K-means clustering algorithm based on rough set
Chen Fuji · Engineering Journal of Wuhan University · 2011
Based on the rough set theory,a new K-means clustering algorithm is proposed.The data depends on the consistency of condition attributes and decision attributes in the decision table.Attribute reduction in rough set is used firstly to eliminate redundant property.Then the weight of left attributes is given according to the attribute significance.Based on above,the improved K-means clustering algorithm is used to analyze the condition attributes.The method's advantage exists in eliminating unimportant attributes and giving out each attribute's weight,which makes clustering more effective and more objective.The experimental results show that this algorithm is efficient.