Plot a correlation matrix. The value of the correlation will be color-coded and additionally printed onto each cell of the 2D plot.
# correlation plots with 5% cutoff
+PlotCorrelations.asimov.medium {
+HWWRun2GGF {
<outputFile="workspaces/run2-ggf-$(fitLabel)/correlations-asimov-medium.pdf">
}
}
+PlotCorrelations.observed.medium {
+HWWRun2GGF {
<outputFile="workspaces/run2-ggf-$(fitLabel)/correlations-observed-medium.pdf">
}
}
@PlotCorrelations.*.medium/? {
<blacklist={"mu"}>
<whitelist={"gamma_*"}>
<cutoff=0.05>
<style.showNumeric=true>
<style.numeric.textSize=0.25>
<style.textSize=0.01>
}
The name of the output file can be determined with the outputFile
tag.
The source of the correlation matrix is identified with the fitresult
tag.
Some additional options exist to control the styling:
orientation="up"/"down"
choose whether the diagonal should be
pointing up or downblacklist
: List expressions matching parameters that should
not be includedwhitelist
: List expressions matching parameters that should be
included, even if also matching the blacklistcutoff
: Numeric value of the minimum correlation to be
plotted. All rows and columns not including any entry above this
value will be dropped.writeCorrelations.outputFile="filename.txt"
,
writeCorrelations.threshold=0.6
: Write all entries of the
correlation matrix exceeding a given threshold to a log file
for further studies.style.showNumeric
Set to false to suppress numbers printed on
cells. The text size of the numbers can be controlled with
style.numeric.textSize
.style.textSize
: Set the text size of everything on the canvas.outputFile.csv
, outputFile.plain
: Give filenames to also
print the correlation matrix to a CSV or plain text file.style.canvasScale
: Control the scale of the canvas.style.color.divisions
: number of steps for the Z (color) axisstyle.color.textSize
: font size for the labels ofthe Z (color) axis style.marginScale
,style.marginScale.top
,style.marginScale.bottom
,style.marginScale.left
,style.marginScale.right
: Control the scale of the margins.