h_reduce: Reducing Parameter Space

In bbolker/cpcbp: common principal components/back-projection analysis

Description

Tests whether to reject the NH that the last p-q eigenvalues are very small and approximately zero variance, based on approximation 4.3 (Flury, 1986).

Usage

  h_reduce(matList, n, q, f, alpha = 0.05)

Arguments

matList	list of covariance matrices
n	numeric vector of sample sizes
q	number of eigenvectors to be kept
f	a fraction value between 0 and 1 which determines the relative contribution of an eigenvalue to CPC
alpha	confidence level

References

Flury B.N. Asymptotic theory for common principal component analysis. Annals Statist. 1986. 14, 418-430.

bbolker/cpcbp documentation built on May 11, 2019, 9:28 p.m.