initialize.x: Diagnose and initialize light level estimation.

initialize.xR Documentation

Diagnose and initialize light level estimation.

Description

Primarily for the purposes of initializing the estimation, these functions can also be used for diagnostic purposes. position.logp produces grids of simplistic position likelihood for each twilight and uses those to initialize positions for running estimations.

Usage

position.logp(model, x1, x2, xrest = NULL, subset = 1:model$n,
initialize.x = TRUE, start = NULL, end = NULL, prob = 0.8, winoffset = 5)

initialize.x(model, x1, x2, xrest = NULL)

light.quantile(model, chain, day, seg, probl = c(0.025, 0.5, 0.975))

show.segment(model, chain, segment, day, light, k, n = 50, ...)

Arguments

model

estimation model object

x1

vector of x-coordinates defining the prior grid

x2

vector of y-coordinates defining the prior grid

xrest

value for remaining parameters - default is light attenuation

subset

evaluate subset of segments - default uses all

initialize.x

logical - create initial points for x?

prob

probability - threshold to apply to overlapping quantiles, defaults to 0.8

winoffset

an odd-numbered window size to use when intersecting subseqent segments - defaults to 5

chain

chain object from estimation

day

POSIXct vector of date-times

seg

desired segment

probl

probability level for quantile

start

known position of release

end

known position of recapture

segment

vector of segment data

light

vector of light data

k

desired segment to show

n

length of vector to evaluate

...

additional arguments to be passed to plot

Details

The primary function here is position.logp, for initializing the estimation for solar.model and metropolis0.

Value

initialize.x returns a matrix with 3 columns, lon,lat,attenuation position.logp returns a list with model running components show.segment is used for its side effect, a plot of light level for a twilight segment light.quantile returns a numeric vector

Author(s)

Michael D. Sumner


Trackage/tripEstimation documentation built on April 24, 2023, 6:57 p.m.