fit_rw: Random Walk with time-varying autocorrelation

Description Usage Arguments Details Value

View source: R/fit_rw.r

Description

Random Walk with time-varying autocorrelation

Usage

1
2
3
fit_rw(data, tstep = 24/24, formula = ~1, center = FALSE, nu = 10,
  g0 = 0.5, span = 0.4, optim = c("nlminb", "optim"), verbose = FALSE,
  amf = amfCRAWL())

Arguments

data

a dataframe of observations (see details)

tstep

time step as a proportion of 24 h

formula

regression formula for the covariates; default: ~ 1 (no covariates)

center

center the covariates; default = FALSE

nu

fixed df for t-distributed measurement errors

g0

initial value for stationary gamma (move autocorrelation) on 0, 1

span

degree of loess smoothing for location initial values

optim

numerical optimizer

verbose

report progress during minimization

amfCRAWL

Argos error class multiplication factors

Details

Fit a random walk with time-varying autocorelation, via TMB, to ssm-filtered track data.

The input track is given as a data_frame where each row is a location 'observed' without error and columns

'date'

observation time (POSIXct, GMT),

'lon'

observed longitude,

'lat'

observed latitude,

Value

a list with components

fitted

a dataframe of estimated states: gamma (tva model); lon, lat, gamma (tvae model)

par

model parameter summmary

data

input data

tmb

the TMB object

opt

the object returned by the optimizer

aic

AIC statistic: 2 * NLL + 2 * k


ianjonsen/bssm documentation built on July 3, 2017, 10:33 p.m.