Abhishek Tiwary wrote:
> with the work. So basically we have the results with two sets of mean
> values and accuracy ranges shown using error bars (see the word
> ducument file). Now I want to use a technique similar to what was
> done in the attached figure to represent the result as a shaded
> region (covering the accuracy range) rather than the conventional
> line with y-error bars plotted in Excel.
That plot style is known as an "error band". gnuplot until version 4.0
could not possibly support it. Nowadays, it might actually be able to,
but you have to lend it a hand. You'll have to plot 3 datasets in
different styles:
1) the experimental data and their errors "with errorbars"
2) the model central value "with lines"
3) the model error region "with filledcurves", and with a specially
crafted data file.
Otherwise, you'll have to resort to plotting the model "with
boxxyerrorbars", or by doing two additional plots "with lines" at the
upper and lower error limit (and maybe fill in that area in
postprocessing, possibly even in CorelDraw if you absolutely have to).
|