fit_ssm: Function for fitting state-space movement model

In PaulRegular/motus: State-space models for animal movement collected using VEMCO telemetry

Function for fitting state-space movement model

Description

Function for fitting state-space movement model

Usage

fit_ssm(track, formula = NULL, gamma_model = "RW",
  gamma_threshold = 0.8, gamma_prob = 0.95, dist = "t",
  silent = FALSE, gr_threshold = 10, start_par = NULL,
  scale = "sd")

Arguments

track	data.frame of an individual track containing time, UTM coordinates, time differences and estimates of observation error. UTM coordinate columns should be labeled lon and lat and units should equal km. time difference should be a ratio of the minimum sampling interval and named delta_t. Observation error should be labeled 1) sd_lon and sd_lat for the normal distribution; 2) tau_lon, tau_lat, nu_lon, and nu_lat for the parameters for the t-distribution; and, 3) scale_lon and scale_lat for the cauchy distribution.
formula	Formula describing relationship between gamma (parameter controlling autocorrelation in the movement process) and covariates. Covariates are not used if set to NULL
gamma_model	Should the gamma parameter (autocorrelation for movement) be "fixed", modeled as a random walk ("RW") or an "AR1" process?
gamma_threshold	A threshold for defining directed or area-restricted phases of movement
dist	Distribution to use for observation error ("null", "normal", "t", "cauchy", "robust"). The "robust" option is a mixture of the normal and t distributions. If NULL, locations are to be "true".
silent	Disable tracing information?
gr_threshold	Stop if maximum gradient exceeds this value (large values indicate convergence issues)
start_par	List of start parameters from a simpler fit from this model (par_est in the list). It may be helpful to fit a simpler version of this model, then supply starting parameters that are closer to where they should be.
scale	Method for scaling x values to improve parameter scaling and aid convergence, where "sd" scales using sd, "max" scales to max deviation, or provide a value.
gamma_prop	Probability of being considered above or below gamma_threshold value (i.e. "directed" or "area-restricted")

PaulRegular/motus documentation built on Jan. 28, 2024, 6:11 p.m.