Build_JTree {treelet}R Documentation

Build a hierarchical/wavelet tree of 2-by-2 Jacobi rotations

Description

Builds a hierarchical binary tree using the covariance structure of the inputs. Decorrelates groups of two with PCA .

Usage

Build_JTree(C, cc, maxlevel)

Arguments

C input covariance matrix
cc correlation coeffs
maxlevel default: dim-1 (levels L=0,1,2,...,maxlevel)

Value

~Describe the value returned

Zpos positions of the two clusters merged; number between 1 and d (the length of the comp vector)
PCidx ID of prinicipal components; [1 2] or [2 1]
T Element T{k} in the list array contains a 2-by-2 rot matrix
all_nodes node labels
theta rotation angles that decorrelates inputs X(:,Z(k,1)) and X(:,Z(k,2))
Z contain the indices of clusters that are merged
all_d labels of the d-components in the comp vector (=0 if s-component, otherwise 1,...,d-1)
C updated covariance matrix
cc updated coefficients

Author(s)

Ann Lee, Di Liu


[Package treelet version 0.1-0 Index]