%0 Journal Article
%T 基于置换检验的聚类结果评估
%A 谷飞洋
%A 田博
%A 张思萌
%A 陈征
%A 何增有
%J 智能系统学报
%D 2016
%R 10.11992/tis.201603038
%X 对聚类结果,传统的评估方法不能从统计意义上对结果评估。ECP是一种新颖的基于置换检验的评估算法。ECP直接对聚类结果进行置换检验从而计算出p-value。为了测试ECP的效果,利用了UCI中的iris, wine, yeast数据集对算法进行评测。实验结果表明,ECP可以在能够接受的时间内运算出比较准确的实验结果。</br>For the result of clustering, tranditional methods of evalution couldn’t assess the result in statistics. We propose a new algorithm called ECP(Statistical evaluation of Clustering based on Permutation test) which uses permutation test to evaluate the result of clustering. To evaluate the performance of the algorithm, we use the data sets, iris, wine, yeast, from UCI datasets. Experimental results show that the performance of the algorithm is good
%K 聚类
%K 聚类评估
%K 统计检验
%K 置换检验&lt
%K /br&gt
%K clustering
%K clustering evaluation
%K statistical test
%K permutation test
%U http://tis.hrbeu.edu.cn/oa/darticle.aspx?type=view&id=20160304