基于时间因子的混沌粒子群优化K-means算法

The K-means algorithm of chaotic particle swarm optimization with time factor

针对传统的K-means算法对初始聚类中心取值敏感和易陷入局部最优解等缺点,提出一种带时间因子的改进粒子群优化(Particle Swarm Optimization,PSO)聚类算法。首先在PSO算法中引入反映时间效应的动态调整时间因子,以避免粒子在最优解附近震荡,为保证粒子在规定范围内运动,采用边界缓冲墙对越界粒子进行处理;其次针对粒子群算法存在的全局搜索性能问题,通过改进的混沌技术对粒子群进行扰动,以混沌搜索替代随机搜索,确保种群的多样性,进而使粒子群向更优的方向移动;最后将改进后的粒子群算法结合K-means算法,以提高粒子的局部勘探能力,从而更快地找到全局最优位置。对UCI中的Iris数据集和Wine数据集仿真表明,该算法相比其他2种算法,聚类准确率分别增长了5.1%和1.3%,1.79%和1.09%。

To solve the shortcomings of the traditional K-means algorithm which is sensitive to the initial clustering centers and easy to fall into local optimum,the K-means algorithm of chaotic-particle swarm optimization with time factor ( KCPTF) is proposed. Firstly, a dynamic adjustment time factor is introduced into the PSO which keeps the particles from shocking around the optimum. Then,in view of the PSO global searching performance problems the improved chaotic technology is used to disturb the particle swarm, replace random search with chaotic search to ensure the particle swarm move to the better direction. Finally,the improved PSO algorithm is combined with K-means algorithm to improve the particles local exploration ability and find the global optimal position quickly. Compared to the other two kinds of algorithms, the simulation results show that KCPTF algorithm accuracy respectively increases 5. 1% and 1.3% ,1.79% and 1.09%.