Statistically validated hierarchical clustering: Nested partitions in hierarchical trees

We develop an algorithm that is fast and scalable in the detection of a nested partition extracted from dendrogram obtained hierarchical clustering multivariate series. Our provides p-value for each clade observed tree. The by computing many bootstrap replicas dissimilarity matrix performing statistical test on difference between associated with given its parent node. prove efficacy our set benchmarks generated hierarchically factor model. compare results those Pvclust. Pvclust widely-used pursuing global approach originally developed context phylogenetic studies. In numerical experiments, we focus role multiple hypothesis correction robustness algorithms to inaccuracies errors datasets. verify much faster than has better scalability both number elements records investigated set. also apply two empirical datasets, one related biological complex system other financial time-series. clusters detected methodology are meaningful respect some consensus partitioning