This style of plot was officially named a "Piner Plot" at the CAPAM Selectivity Workshop, La Jolla March 2013. This is in honor of Kevin Piner's contributions to interpreting likelihood profiles. He's surely not the first person to make such a plot but the name seems to have stuck.
PinerPlot(
summaryoutput,
plot = TRUE,
print = FALSE,
component = "Length_like",
main = "Changes in length-composition likelihoods by fleet",
models = "all",
fleets = "all",
fleetnames = "default",
profile.string = "R0",
profile.label = expression(log(italic(R)[0])),
exact = FALSE,
ylab = "Change in -log-likelihood",
col = "default",
pch = "default",
lty = 1,
lty.total = 1,
lwd = 2,
lwd.total = 3,
cex = 1,
cex.total = 1.5,
xlim = "default",
ymax = "default",
xaxs = "r",
yaxs = "r",
type = "o",
legend = TRUE,
legendloc = "topright",
pwidth = 6.5,
pheight = 5,
punits = "in",
res = 300,
ptsize = 10,
cex.main = 1,
plotdir = NULL,
add_cutoff = FALSE,
cutoff_prob = 0.95,
verbose = TRUE,
fleetgroups = NULL,
likelihood_type = "raw_times_lambda",
minfraction = 0.01
)
List created by the function
SSsummarize()
.
Plot to active plot device?
Print to PNG files?
Which likelihood component to plot. Default is "Length_like".
Title for plot. Should match component.
Optional subset of the models described in
summaryoutput
. Either "all" or a vector of numbers indicating
columns in summary tables.
Optional vector of fleet numbers to include.
Optional character vector of names for each fleet.
Character string used to find parameter over which the
profile was conducted. If exact=FALSE
, this can be a substring of
one of the SS parameter labels found in the Report.sso file.
For instance, the default input 'R0'
matches the parameter 'SR_LN(R0)'. If exact=TRUE
, then
profile.string needs to be an exact match to the parameter label.
Label for x-axis describing the parameter over which the profile was conducted.
Should the profile.string
have to match the parameter
label exactly, or is a substring OK.
Label for y-axis. Default is "Change in -log-likelihood".
Optional vector of colors for each line.
Optional vector of plot characters for the points.
Line total for the likelihood components.
Line type for the total likelihood.
Line width for the likelihood components.
Line width for the total likelihood.
Character expansion for the points representing the likelihood components.
Character expansion for the points representing the total likelihood.
Range for x-axis. Change in likelihood is calculated relative to values within this range.
Maximum y-value. Default is 10\ plotted.
The style of axis interval calculation to be used for the x-axis (see ?par for more info)
The style of axis interval calculation to be used for the y-axis (see ?par for more info).
Line type (see ?plot for more info).
Include legend?
Location of legend (see ?legend for more info).
Width of plot
Height of plot
Units for PNG file
Resolution of plots printed to files.
The default is res = 300
.
Point size for PNG file
Character expansion for plot titles
Directory where PNG files will be written. by default it will be the directory where the model was run.
Add dashed line at ~1.92 to indicate 95% confidence interval
based on common cutoff of half of chi-squared of p=.95 with 1 degree of
freedom: 0.5*qchisq(p=cutoff_prob, df=1)
. The probability value
can be adjusted using the cutoff_prob
below.
Probability associated with add_cutoff
above.
Return updates of function progress to the R GUI? (Doesn't do anything yet.)
Optional character vector, with length equal to the number of declared fleets, where fleets with the same value are aggregated
choice of "raw" or "raw_times_lambda" (the default) determines whether or not likelihoods plotted are adjusted by lambdas (likelihood weights)
Minimum change in likelihood (over range considered) as a fraction of change in total likelihood for a component to be included in the figure.
Kevin Piner says that he's not the originator of this idea so Athol Whitten is going to add a reference here.