基于微正则退火算法对K-means聚类算法的优化

K-means算法是经典的基于划分的聚类算法，但该算法存在依赖于初始聚类中心、容易陷入局部最优解等缺点，针对这些缺点，本文提出了基于微正则退火K-means聚类算法，通过继承微正则退火算法的高效全局寻优特性，可以避免陷入局部最优解。实验结果表明，改进的算法能够有效的减少原算法对初始聚类中心点的依赖，提高算法的稳定性，摆脱原算法容易陷入局部最优解的缺点。

Optimization for K-means Clustering Algorithm Based on Microcanonical Annealing Algorithm

K-means algorithm is one of the clustering algorithm based on partition. But the K-means clustering algorithm is dependent on the initial clustering center and easy to fall into local optimal solutions, Therefore, an improved K-means clustering algorithm based on the Microcanonical annealing algorithm is put forward, which has efficient global optimization characteristics, and can avoid falling into local optimal solutions. The experimental results show that the improved algorithm can effectively reduce the dependence of the original algorithm on the initial clustering center, and improve the stability of the algorithm, getting rid of the defect that the original algorithm is easy to fall into local optimal solutions.