Confidence intervals/significance testing for dendrogram

Question

Matt 2013-9-8

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/86644-confidence-intervals-significance-testing-for-dendrogram

回答： Sebastien De Landtsheer 2018-2-1

Is there a simple and/or straightforward way to generate confidence intervals or significance measures for each of the branches when using the dendrogram function in Matlab?

I am just using the standard code for generating the dendrogram:

Y=pdist(Matrix, 'euclidean'); Z=linkage(Y,'average'); [H,T]=dendrogram(z,20);

and would like to assess the clustering quality. From what I can tell, it would require some permutation-based inferences.

I was hoping for a built-in function, but can't find one.

There was a relatively recent publication doing this for gene research: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3023458/

"we introduce a permutation test based on comparing the within-cluster structure of the observed data with those of sample datasets obtained by permuting the cluster membership. We carry out this test at each node of the dendrogram using a statistic derived from the singular value decomposition of variance matrices. The p-values thus obtained provide insight into the significance of each cluster division"

Could this be easily implemented?

Thanks

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Sebastien De Landtsheer 2018-2-1

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/86644-confidence-intervals-significance-testing-for-dendrogram#answer_302952

As far as I know, the standard way to assess support for nodes in a tree is to compute multiple trees from bootstrapped data and counting how often a given group of the original tree appears in the perturbed ones, then indicating the support on the original tree. I have some code doing that, I might write a function eventually.