sumFunc {ebdbNet} | R Documentation |
Helper function for hankel
to determine what percentage of variance is explained by the singular
values of the block-Hankel matrix.
sumFunc(x, cutoff)
x |
Vector of singular values from singular value decomposition of block-Hankel matrix |
cutoff |
Value to determine cutoff to be considered for singular values (e.g., 0.90) |
This function is used to help determine the number of large singular values, which in turn determines the hidden state dimension to be estimated.
The number of large singular values, as defined by the user-supplied cutoff.
Andrea Rau
library(ebdbNet) tmp <- runif(1) ## Initialize random number generator set.seed(1241901) ## Set seed x <- matrix(rnorm(100), nrow = 10, ncol = 10) svd <- svd(x)$d dim <- sumFunc(svd, 0.90) ## 7 singular values contribute 90