dot-VAR: (Fast) Barebones Vector-Autoregression

In dfms: Dynamic Factor Models

(Fast) Barebones Vector-Autoregression

Description

Quickly estimate a VAR(p) model using Armadillo's inverse function.

Usage

.VAR(x, p = 1L)

Arguments

x	data numeric matrix with time series in columns - without missing values.
p	positive integer. The lag order of the VAR.

Value

A list containing matrices Y = x[-(1:p), ], X which contains lags 1 - p of x combined column-wise, A which is the np \times n transition matrix, where n is the number of series in x, and the VAR residual matrix res = Y - X %*% A.

A list with the following elements:

Y	x[-(1:p), ].
X	lags 1 - p of x combined column-wise.
A	np \times n transition matrix, where n is the number of series in x.
res	VAR residual matrix: Y - X %*% A.

Examples

var = .VAR(diff(EuStockMarkets), 3)
str(var)
var$A
rm(var)

dfms documentation built on June 22, 2024, 10:31 a.m.