结合遗传k均值改进的密度峰值聚类算法

针对密度峰值聚类(CFSFDP)算法处理多密度峰值数据集时,人工选择聚类中心易造成簇的误划分问题,提出一种结合遗传k均值改进的密度峰值聚类算法。在CFSFDP求得的可能簇中心中,利用基于可变染色体长度编码的遗传k均值的全局搜索能力自动搜索出最优聚类中心,同时自适应确定遗传k均值的交叉概率,避免早熟问题的出现。在UCI数据集上的实验结果表明,改进算法具有较好的聚类质量和较少的迭代次数,验证了所提算法的可行性和有效性。

Improved density peak clustering algorithm combining genetic k-means

For the clustering by fast search and find of density peaks(CFSFDP),when dealing with multi-density peak data sets,the manual selection of clustering centers is easy to cause cluster misclassification.An improved density peak clustering algorithm combining genetic k-means was proposed.The global search ability of genetic k-means based on the modified variable string length was used to automatically search for the optimal cluster centers from the possible cluster centers obtained by using CFSFDP.The genetic k-means adaptively determined the crossover probability to avoid the emergence of premature problems.The results of experiment on the UCI data set show that the improved algorithm has better clustering quality and less iteration,which verifies the feasibility and effectiveness of the proposed algorithm.