rr.regression: Reduced rank regression

In b-brune/tvRRR: Time-varying reduced-rank regression using state-space models

Description

RRR with additional full-rank predictors taken from Reinsel and Velu (1998), Ch. 3, Thm. 3.1 u corresponds to their z. The model is

Usage

rr.regression(X, y, u = NULL, rank, Gamma_type = "identity")

Arguments

X	matrix of predictors, dimensions (t x q)
y	matrix of responses, dimensions (t x p)
u	matrix of additional predictors with full rank coefficient matrix, dimensions (t x k), optional.
rank	presumed rank of the coefficient matrix
Gamma_type	normalization, can be "identity", "OLS" or "cov_y", see Reinsel and Velu for details

Details

y = mu + A B X + D u + epsilon

Value

A named list with elements A, B, C = A %*% B, D and mu.

Model: y = mu + ABX + Du + error

b-brune/tvRRR documentation built on Dec. 19, 2021, 6:37 a.m.