一种改进K-means聚类的FCMM算法

Algorithm named FCMM to improve K-means clustering algorithm

针对K-means算法易受初始聚类中心影响而陷入局部最优的问题,提出一种基于萤火虫智能优化和混沌理论的FCMM算法。利用最大最小距离算法确定聚类类别值K和初始聚类中心位置,以各聚类中心为基准点,利用Tent映射构建混沌空间,通过混沌搜索更新聚类中心,以降低初始聚类中心过于临近的影响,并改善算法易陷入局部最优的问题。仿真结果表明,FCMM算法的平均聚类精度相较于经典K-means算法和FA算法分别提高了7.51%和2.2%,成功避免算法陷入局部最优解,提高了划分初始数据集的效率和寻优精度。

In order to solve the problem that the K-means algorithm gets affected by the initial cluster centers easily, this paper proposed FCMM algorithm based on firefly intelligence optimization and chaos theory. It used the max-min distance clustering algorithm to calculate the number K of cluster center and determined the location of initial cluster centers. To overcome the problem that initial clustering centers are too close to each other and traditional algorithm falls into local optima easily, this algorthm used Tent mapping to construct a chaotic space with each cluster center as the datum point, and then updated cluster centers through chaotic search. The experimental results show that the average clustering accuracy of the FCMM algorithm than that of the classical K-means algorithm and the FA algorithm is respectively 7.51% and 2.2% higher, the FCMM algorithm avoids falling into the local optimal solution successfully, and improves the efficiency and precision of the initial data set.