基于核函数的层次聚类算法

A hierarchical clustering algorithm based on kernel function

层次聚类算法是运行复杂度较高的聚类算法,基于不相似性测度的层次聚类算法不适合稀疏高维数据.结合核函数特点,提出了一种基于核函数的层次聚类算法.利用该算法,对稀疏高维数据进行了层次聚类对比,实验结果表明,该算法提高了层次聚类的准确率.

Hierarchical clustering algorithm is a running complexity higher clustering algorithm,hierarchical clustering algorithm based on non-similarity measure is not suitable for high-dimensional sparse data.Combining features of kernel functions,A hierarchical clustering algorithm based on kernel function is raised.Using this algorithm,comparing hierarchical clustering results with high-dimensional sparse data,the experimental results show that the algorithm can improve the accuracy of hierarchical clustering.