Plotting Pareto fronts with plot_pf()#

Various examples using mooplot.plot_pf().

Plot 2D datasets#

The default is type='points'.

import moocore
import mooplot

sets = moocore.read_datasets(moocore.get_dataset_path("input1.dat"))
# Select only the first 5 datasets
sets = sets[sets[:, -1] <= 5, :]
fig = mooplot.plot_pf(sets, title="input1.dat")
fig


Plot a two objective dataset with points and lines

fig = mooplot.plot_pf(sets, type="p,l")
fig


Plot 3D datasets#

filename = moocore.get_dataset_path("spherical-250-10-3d.txt")
sets = moocore.read_datasets(filename)
# Select only the last 3 datasets
sets = sets[sets[:, -1] >= 8, :]
fig = mooplot.plot_pf(sets, title="Spherical")
fig


Three objective surface#

Use the type="surface" argument to draw a smooth 3D surface for each dataset.

fig = mooplot.plot_pf(sets, type="surface")
fig


Three objective surface+points#

Use type="surface, points" to add both points and surfaces to the plot. You can compare the surface of one dataset to the points of another by clicking on “set 1” and “set 2 points” in the legend to hide some of the points and surfaces.

filename = moocore.get_dataset_path("uniform-250-10-3d.txt")
sets = moocore.read_datasets(filename)
sets = sets[sets[:, -1] >= 8, :]
fig = mooplot.plot_pf(sets, type="surface, points", title="Uniform")
fig


Three objective cube graph#

Use type="cube" to add a cuboid for each point.

Warning

This may be slow in large datasets.

fig = mooplot.plot_pf(sets, type="cube")
fig


Total running time of the script: (0 minutes 0.399 seconds)

Gallery generated by Sphinx-Gallery