PSM.smooth {PSM}R Documentation

Smoothing of model states based on estimated population parameters.

Description

Gives estimates of model states and random effects eta. The function is intended to be used based on population parameters found using PSM.estimate or to check initial values before parameter estimation.

Usage

PSM.smooth(Model, Data, THETA, subsample = 0, trace = 0, etaList = NULL)

Arguments

Model Model list.*
Data Data list.*
THETA Vector of population parameters used for the state estimation.
subsample Number of points to estimate states in between measurements. The extra points are linearly spaced.
trace Non-negative integer. If positive, tracing information on the progress of the optimization is produced. Higher values produces more tracing information.
etaList Matrix where each column contains an etimate of eta_i. etaList has the same format as the output of PSM.estimate. If ommitted, the function will evalutate the population likehood function to find estimates of eta for all individuals.

* See description in PSM.estimate.

Details

The function produces three types of estimates.

Predicted
Only past measurements are used for the state estimate at time t.
Filtered
Only past and the current measurements are used for the state estimate at time t.
Smoothed
All measurements (both past and future) are used to form the state estimate at time t. This is usually the prefered type of state estimate.

If subsample>0 then the data is automatically subsampled to provide estimated of the model states between observation time points.

Value

An unnamed list with one element for each individual. Each element contains the following elements:

Time Possibly subsampled time-vector corresponding to the estimated states
Xs, Ps Smoothed state and state co-variance estimate
Ys Response based on smoothed state: Ys = g(Xs).
Xf, Pf Filtered state and state co-variance estimate
Xp, Pp Predicted state and state co-variance estimate
Yp, R Predicted observations and observation variances
eta Estimated eta
negLogL Value of the negative log-likelihood function at THETA (thus same value for all individuals).

Note

For further details please also read the package vignette pdf-document by writing vignette("PSM") in R.

Author(s)

Stig B. Mortensen and Søren Klim

References

Please visit http://www.imm.dtu.dk/psm or refer to the help page for PSM.

See Also

PSM, PSM.estimate, PSM.simulate, PSM.plot, PSM.template

Examples

cat("\nExamples are included in the package vignette.\n")
[Package PSM version 0.8-3 Index]