eigenPower: Function eigenPower.
In variani/cpca: Methods to perform Common Principal Component Analysis (CPCA)

Description Usage Arguments Value Note Examples

View source: R/eigen.lib.R

The function implements the power algorithm for EVD.

Function eigenPowerRcpp.

Function eigenPowerRcppEigen.

Function eigenPowerRcppParallel.

Function eigenPowerEigenParallel.

eigenPower(A, v0, tol = 1e-06, maxit = 1000, sparse = FALSE,
  sparseSymm = FALSE, ncomp = 1, verbose = 0)

eigenPowerRcpp(A, v0, tol = 1e-06, maxit = 1000, mode = 1, verbose = 0)

eigenPowerRcppEigen(A, v0, tol = 1e-06, maxit = 1000, ncomp = 1,
  symmetric = FALSE, verbose = 0)

eigenPowerRcppParallel(A, v0, tol = 1e-06, maxit = 1000, cores = -1,
  chunkSize = 1, verbose = 0)

eigenPowerEigenParallel(A, v0, tol = 1e-06, maxit = 1000, cores = -1,
  chunkSize = 1, verbose = 0)

`A`	A two-dimensional square matrix, either of `matrix` or `Matrix` class.
`v0`	A numeric vector; the initial guess for eignevector. If it is missing, a random vector is generated.
`tol`	The tolerance threshold used to stop when reaching no improvement if estmiation of eigenvalue. The default value is `1e-6`.
`maxit`	The maximum number of iterations. The default value is `1e4`.
`sparse`	The boolean value, whether to convert the input `A` matrix to one of the `Matrix` classes. The default value is `FALSE`.
`sparseSymm`	The boolean value, whether to convert the input `A` matrix to one of the `Matrix` classes, while trying to convert to a symmetric matrix type. The default value is `FALSE`.
`ncomp`	The number of eigenvectors to be extracted. The default value is `1`. The value of `0` means extract all eigenvectors.
`verbose`	The integer value indicating the verbose level. The default value is `0`.
`mode`	An integer indicating the mode of implementation (for `eigenPowerRcpp` function). The default value is `1`.
`symmetric`	A logical which value says explicetly if the input matrix `A` is symmetric. The default value is `FALSE`.
`cores`	The number of cores (for parallel versions). The default value is `-1`. This argument is passed next to `RcppParallel::setThreadOptions(numThreads = cores)`.
`chunkSize`	The minimal size of a chunk (for parallel versions). The default value is `1`. This argument is passed next to a wrapper `eigenPower_Rcpp_Parallel`.

A list several slots: v the first eigenvector; lambda the first eigenvalue; etc.

This function is inspired by the post http://blogs.sas.com/content/iml/2012/05/09/the-power-method.html.

A <- matrix(c(-261, 209, -49, 
  -530, 422, -98,
  -800, 631, -144),
  ncol = 3, nrow = 3, byrow = TRUE)
  
v0 <- c(1, 2, 3)

out <- eigenPower(A, v0, maxit = 40, verbose = 2)

out[c("lambda", "v")]