mom_fit: Fit the reduced dimension covariance matrix

In jelsema/RRSM: Methods for reduced rank spatial models

Description

Uses the binned empirical covariance matrix for Method of Moments (MOM) estimation to fit the reduced dimension covariance matrix of a reduced rank spatial model.

Usage

mom_fit(fitting = "frobenius", SigM, Sbar, mcn = 1000, vlift = "none",
  ...)

Arguments

fitting	string describing which fitting method to be used to fit the covariance matrix. See details.
SigM	the binned empirical covariance matrix.
Sbar	a similarly binned version of the matrix of basis functions.
mcn	if positive, winsorizes eigenvalues of covariance matrix so that the condition number is at max mcn.
vlift	if TRUE, lifts eigenvalues of robust covariance matrix ("proportional").
...	space for additional arguments to be passed to the fitting function.

Details

The Method of Moments (MOM, method="MOM") estimation is a two-step procedure. First, an empirical binned covariance matrix is computed (see mom_bec), and then the covariance matrix is fitted. Different options for the fitting stage can be selected using fitting. Currently there are two choices:

• "frobenius"The covariance is fitted by minimizing the Frobenius norm (REF).

• "robust"The covariance is fitted by minimizing a robust norm (using quantile regression) (REF).

Value

A matrix.

Note

The values of SigM and Sbar are intended to be the output values of mom_bec

See Also

rr_est mom_bec

jelsema/RRSM documentation built on May 19, 2019, 4:02 a.m.