kalman: Run a Kalman filter

Description Usage Arguments Value

View source: R/wrap.r

Description

Function to run a Kalman filter on the sde approximation of a ssm.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
kalman(
  ssm,
  dt = NULL,
  id = 0,
  root = NULL,
  n_obs = NULL,
  eps_abs_integ = NULL,
  eps_rel_integ = NULL,
  like_min = NULL,
  freeze_forcing = NULL,
  interpolator = NULL,
  verbose = FALSE,
  warning = FALSE,
  no_dem_sto = FALSE,
  no_white_noise = FALSE,
  no_diff = FALSE,
  traj = TRUE,
  hat = FALSE,
  trace = TRUE,
  diag = TRUE,
  prior = FALSE,
  seed_time = TRUE
)

Arguments

ssm

a ssm object, returned by new_ssm.

dt

numeric, integration time step.

id

integer, unique integer identifier that will be appended to the output.

root

character, root path for output files (if any) (no trailing slash). If NULL (default), outputs are written in "your_model_path/the_name_of_the_wrapper".

n_obs

numeric, number of observations to be fitted (for tempering). If NULL (default), all observations are fitted.

eps_abs_integ

numeric, absolute error for adaptive step-size control.

eps_rel_integ

numeric, relative error for adaptive step-size control.

like_min

numeric, particles with likelihood smaller than like_min are considered lost. If NULL (default) lower bound on likelihood based on machine precision.

freeze_forcing

character, freeze covariates to their value at specified date (in YYYY-MM-DD format).

interpolator

character, gsl interpolator for covariates

verbose

logical, print logs (verbose). Default to FALSE.

warning

logical, print warnings. Default to FALSE.

no_dem_sto

logical, turn off demographic stochasticity (if any). Default to FALSE.

no_white_noise

logical, turn off white noises (if any). Default to FALSE.

no_diff

logical, turn off diffusions (if any). Default to FALSE.

traj

logical, print the trajectories. Default to TRUE.

hat

logical, print the state estimates. Default to FALSE.

trace

logical, print the trace. Default to TRUE.

diag

logical, print the diagnostics outputs (e.g. prediction residuals). Default to TRUE.

prior

logical, add log(prior) to the estimated log-likelihood. Default to TRUE.

seed_time

logical, seed the random number generator with the current time. Default to TRUE.

Value

a ssm object updated with latest SSM output and ready to be piped into another SSM block.


StateSpaceModels/ssminr documentation built on Feb. 7, 2020, 8:20 p.m.