EstimSigma {chickn} | R Documentation |
The mean data variance estimation.
EstimSigma( Data, ind.col, m, nblocks = 32, niter = 3, sigma_start = 0.1, nparts = 1, ... )
Data |
A Filebacked Big Matrix n x N. Data signals are stored in the matrix columns. |
ind.col |
Column indeces for which the data sketch is computed. By default all matrix columns. |
m |
Number of frequency vectors. |
nblocks |
Number of blocks, on which the regression is performed. Default is 32. |
niter |
Number of iterations. Default is 3. |
sigma_start |
An initial value of the data variance. Default is 0.1. |
nparts |
Number of parts to split the data for the data sketch computation. |
... |
Additional arguments passed on to |
The estimated data variance.
The idea of the variance estimation on the data fraction is taken from Keriven N, Bourrier A, Gribonval R, PĂ©rez P (2018). “Sketching for large-scale learning of mixture models.” Information and Inference: A Journal of the IMA, 7(3), 447–508..
DrawFreq
, Sketch
, GenerateFrequencies
X = matrix(rnorm(1e5), ncol=1000, nrow = 100) X_FBM = bigstatsr::FBM(init = X, ncol=1000, nrow = 100) sigma = EstimSigma(Data = X_FBM, ind.col = seq(1,1000, by = 2), m = 20, nblocks = 4)