multModel: Multivariate process models constructor

In smde: Sparse Multivariate Differential Equations

Description

Constructs multivariate models for dynamic processes.

Usage

multModel(y, x, t, suff = TRUE, weights = "unit", sigma = 1)

Arguments

y	a numeric. A response matrix of dimensions d by m.
x	a numeric. A predictor matrix of dimensions d by m.
t	a numeric. A single time step size, or a vector of time step sizes of length m.
suff	a logical. Should sufficient statistics and the MLE be computed.
weights	a character or numeric. Specifies parameter weights. Default value 'unit' means that all parameters are weighted equally with a unit weight, whereas 'adaptive' means that the parameters are weighted inverse proportionally to the MLE (will only be used if suff = TRUE). Alternatively, a numeric specifying the individual parameter weights.

Value

an object of class multModel containing the elements

y	the d by m data matrix
x	the d by m data matrix
t	the time step(s)
ss	sum of squares (computed if suff = TRUE)
xyt	d by d crossproduct of x and y (computed if suff = TRUE)
xxt	d by d crossproduct of x (computed if suff = TRUE)
B	the MLE if suff = TRUE, and the zero matrix otherwise
w	a matrix of parameter weights
d	dimension of the process
m	number of d-dimensional observations
n	equals m * d
p	quals d * d
sigma	the observation variance parameter
status	0 means no problems or errors encountered
msg	error messages or other notable conditions

Author(s)

Niels Richard Hansen [e-mail]

smde documentation built on May 2, 2019, 4:58 p.m.