基于粒子群的三支聚类算法

针对K均值聚类(K-means)算法处理复杂问题时易陷入局部最优值、聚类质量较差等不足，提出一种基于粒子群的三支聚类算法.该算法先以随机产生的聚类中心组合作为初始粒子，构成粒子群；然后，通过调整算法中的速度公式参数，使粒子在迭代过程中能较快速地找出全局最优解，即最优的聚类中心；最后，采用三支决策的方法考察数据与类的关系，把确定归属的数据分配到类的核心域，归属不确定的数据分配到类的边界域.实验结果验证了所提算法的有效性，在寻找全局最优值和聚类结果准确性等方面算法都具有较好的性能.

A three-way clustering algorithm based on particle swarm optimization

In order to solve the problem that K-means algorithm is easy to fall into local optimal value and poor clustering quality when dealing with complex problems, a three-way clustering algorithm based on particle swarm is proposed in this paper. In the proposed algorithm, some combinations of clustering centers are randomly generated as initial particles to form a particle swarm. Then, by adjusting the particle velocity parameters, the global optimal solution can be found quickly in the iterative process, that is, the optimal clustering center can be obtained. Finally, the three-way decision method is used to investigate the relationship between data points and the classes, and the data point with certain attribution is assigned to the core domain of the class, while the data point with uncertain attribution is assigned to the boundary domain of the class. Experimental results verify the effectiveness of the proposed algorithm, which has good performance in finding the global optimal value and the accuracy of the clustering results.