rescale_variance: Re-scales each row of matrix by its MAD estimate

In HDCD: High-Dimensional Changepoint Detection

Re-scales each row of matrix by its MAD estimate

Description

R wrapper for C function computing the (rescaled) median absolute difference in differences for each row of the input matrix. The rescaling factor is set to 1.05 (corresponding to the Normal distribution). Each row of the input matrix then re-scaled by the corresponding noise estimate.

Usage

rescale_variance(X, debug = FALSE)

Arguments

X	A p \times n matrix
debug	If TRUE, diagnostic prints are provided during execution

Value

A list containing

X	the input matrix, variance re-scaled and flattened
scales	vector of MAD estimates of the noise level of each row of the input matrix

Examples

library(HDCD)
n = 200
p = 500
set.seed(101)
# Generating data
X = matrix(rnorm(n*p), ncol = n, nrow=p)

ret = rescale_variance(X)
ret$X #rescaled matrix
ret$scales #estimated noise level for each time series (each row)

# Note that the rescaled matrix is in (column wise) vector form. To transform it back to a matrix,
# do the following:
rescaled_X = matrix(ret$X, nrow = p, ncol=n)

HDCD documentation built on June 22, 2024, 10:53 a.m.