摘要
针对初始聚类中心的选择对于K-均值算法的聚类结果非常敏感,且容易陷入局部极值的缺点,提出利用蚁群聚类算法来搜寻K-均值的初始聚类中心,同时通过在搜索空间增加一组逐渐递减的服从均匀分布的扰动因子,建立基于扰动因子的准则函数下的聚类算法.最后对蚁群聚类算法、K-均值聚类算法以及改进后的算法做了对比实验.实验结果表明,改进后算法的聚类能力更强.
As to the clustering results of K-Means algorithm is very sensitive to selecting an initial cluster centers,and easy to fall into local extreme,it is put forward that K-means＇ initial clustering center is searched by ant colony clustering algorithm which has a strong ability to deal with local extremum.At the same time,clustering algorithm under the criterion function based on the disturbance factors is established by adding a set of gradually decreasing uniform distribution factors in search space.Finally,the contrast tests are made among the ant colony algorithm,the K-mean algorithm and the improved algorithm.The results show that the improved algorithm＇s clustering ability is stronger than the other two.
出处
《纺织高校基础科学学报》
CAS
2017年第1期81-86,共6页
Basic Sciences Journal of Textile Universities
基金
陕西省自然科学基金资助项目(2015JM1012)
关键词
K-均值聚类算法
聚类中心
扰动因子
蚁群聚类算法
K-means algorithm
clustering center
disturbance factor
ant colony clustering algorithm
