Clustering Algorithm for Mixed Attributes Data Based on Restricted Particle Swarm Optimization

摘要

When the data mixed with numerical and categorical values is processed at present,it is very common to convert the categorical values into numerical ones and then cluster them according to a certain weight Obviously such clustering results rely heavily on the weight given by experts.Hence in this paper a categorical attribute is proposed as a potential field restriction to limit the searching directions of particle swarm,which consequently improves the speed and effectiveness of the clustering algorithms.