一种改进的人工鱼群聚类算法

为了克服K-Means算法对初始类簇中心、噪声点、孤立点敏感缺点,将K-Means算法和人工鱼群算法结合,提出了改进的人工鱼群聚类算法。在该算法中将类簇中心看作一条人工鱼,让每条人工鱼执行随机、觅食、聚群、追尾行为中的一种,并将更新后的位置作为K-Means算法的初始值,不断重复人工鱼的位置更新和K-Means操作,直到算法结束。由于在算法中加入了动态移动步长和全局人最优人工鱼位置,聚类的收敛精度和速度都得到提高。使用iris和glass数据集进行聚类时,与其他算法相比,文中的收敛时间缩短2.6%,精度提高1.36%。

An Improved Artificial Fish Swarm Clustering Algorithm

In order to overcome the shortcomings of K-Means of being sensitive to the initial class cluster center,noises,isolated points,this article present an improved artificial fish clustering algorithm,which combines K-Means algorithm and artificial fish swarm algorithm（AFSA）.This algorithm take the center of the cluster as an artificial fish,and allows each fish to randomly select one of the random behaviors,which are preying behavior,swarming behavior and following behavior.The value of the updated location of artificial fish can be seen as the initial value of K-Means.The location updating and K-Means operation are repeateduntil the end of the algorithm.For dynamic moving step and global best artificial fish position is added in this algorithm,clustering convergence accuracy and speed are improved.Compared with other algorithms,the convergence time of this article shortens 2.6％ and the accuracy inereases 1.36 ％ in application of iris and glass datasets clustering,.