groupSparseEigen: Group-wise sparse spectral decomposition of a matrix

Description Usage Arguments Details Value See Also

View source: R/groupSparseEigen.R

Description

Computes group-wise sparse eigenvalues and eigenvectors of a matrix.

Usage

1
2
3
4
5
6
7
8
9
groupSparseEigen(
  x,
  groups,
  y = NULL,
  lambda = NULL,
  kernel = "linear",
  bandwidth = NULL,
  trace = FALSE
)

Arguments

x

numeric matrix; data with variables as columns and samples as rows.

groups

boolean matrix; affiliation matrix. Does the variable in row i pertain to the group in column j?

y

numeric matrix; confounder matrix with rows as samples and columns as confounding factors.

lambda

the tuning parameter, non-negative.

kernel

the kernel to use: "linear", "gaussian".

bandwidth

bandwidth h for Gaussian kernel. Optional.

trace

logical; should progress be printed?

Details

For each group defined in the groups parameter, variables not pertaining to the group see their covariance set to zero. Eigendecomposition is then performed on that matrix.

Value

A list containing the sparse eigenvector and eigenvalues.

See Also

eigen


oliviermfmartin/HelpingHand documentation built on Oct. 10, 2020, 5:59 a.m.