Description Usage Arguments Details Value See Also
Function SSModel
creates a state space object
object of class SSModel
which can be used as an
input object for various functions of KFAS
package.
1 2 3 4 5 6 |
y |
A time series object of class |
Z |
System matrix or array of observation equation. |
H |
Covariance matrix or array of disturbance terms ε[t] of observation equation. Omitted in case of non-Gaussian distributions. Augment the state vector if you want to add additional noise. |
T |
System matrix or array of transition equation. |
R |
System matrix or array of transition equation. |
Q |
Covariance matrix or array of disturbance terms η[t]. |
a1 |
Expected value of the initial state vector α[1]. |
P1 |
Covariance matrix of α[1]. In the diffuse case the non-diffuse part of P[1]. |
P1inf |
Diffuse part of P[1]. Diagonal matrix with ones on diagonal elements which correspond to the unknown initial states. |
u |
Only used with non-Gaussian distribution. See details. |
distribution |
Specify the distribution of the observations. Default is "Gaussian". |
transform |
The functions of |
tolF |
Tolerance parameter for Finf. Smallest value not counted for zero. |
tol0 |
Tolerance parameter for LDL decomposition, determines which diagonal values are counted as zero. |
The custom state space model is constructed by using the
given system matrices Z
, H
, T
,
R
, Q
, a1
, P1
and
P1inf
. Matrix or scalar Z
(array in case of
time-varying Z
) is used to determine the number of
states m. If some of the other elements of the
object are missing, SSModel
uses default values
which are identity matrix for T
, R
(or
k first columns of identity matrix) and
P1inf
, and zero matrix for H
, Q
,
P1
and , a1
. If P1
is given and
P1inf
is not, the it is assumed to be zero matrix.
If Q
is given, it is used to define r, the
dimensions of Q
, which can be smaller than m
(defaults to m).
The linear Gaussian state space model is given by
y[t] = Z[t]α[t] + ε[t], (observation equation)
α[t+1] = T[t]α[t] + R[t]η[t], (transition equation)
where ε[t] ~ N(0,H[t]), η[t] ~ N(0,Q[t]) and α[1] ~ N(a[1],P[1]) independently of each other. In case of non-Gaussian observations, the observation equation is of form p(y[t]|θ[t]) = p(y[t]|Z[t]α[t]), with p(y[t]|θ[t]) being one of the following:
If observations are Poisson distributed, parameter of Poisson distribution is u[t]λ[t] and θ[t]=log(λ[t]).
If observations are from binomial distribution, u is a vector specifying number the of trials at times 1,…,n, and θ[t] = log(π[t]/(1-π[t])), where pi[t] is the probability of success at time t.
For non-Gaussian models u[t]=1 as a default. For Gaussian models, parameter is omitted.
Only univariate observations are supported when observation equation is non-Gaussian.
object of class SSModel
with elements
arimaSSM
for state space representation of
ARIMA model, regSSM
for state space
representation of a regression model, and
structSSM
for structural time series model.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.