期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2015
卷号：8
期号：7
页码：215-224
DOI：10.14257/ijhit.2015.8.7.20
出版社：SERSC
摘要：Normally, improving the performance of clustering depends on improvement of the algorithm. On the basis, this paper presents a hybrid strategy optimization algorithm that K-means algorithm effectively combined with PSO algorithm, which not only has played their respective advantages, but also reflected a hybrid performance. First of all, combined with a semi-supervised clustering idea ， to optimize the clustering center of particle by K - means in the iteration of a lgorithm, enhanced the searching capability of the particles. Secondly, improved the traditional K - means enhance the ability of the algorithm to deal with the concave and convex points. Finally, the algorithm is introduced into the particle state determ ination mechanism, on implementing mutation for unstable particles, so that the algorithm to obtain stable performance. Experimental results show that the hybrid algorithm optimization ability is outstanding, and the convergence and stability can be effectively improved.
