.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "examples/xmpl_ternary_contours.py" .. LINE NUMBERS ARE GIVEN BELOW. .. only:: html .. note:: :class: sphx-glr-download-link-note :ref:`Go to the end ` to download the full example code. or to run this example in your browser via Binder .. rst-class:: sphx-glr-example-title .. _sphx_glr_examples_xmpl_ternary_contours.py: ============================================== Draw contour function of ternary simplex space ============================================== This example illustrates how to draw contourplots for functions with 3 probability inputs and multiple outputs. .. GENERATED FROM PYTHON SOURCE LINES 9-18 .. code-block:: Python # Author: Miquel Perello Nieto # License: new BSD import matplotlib.pyplot as plt import numpy as np np.random.seed(42) print(__doc__) .. GENERATED FROM PYTHON SOURCE LINES 19-21 We show first how to draw a heatmap on a ternary probability simplex, in this case we will define a Dirichlet function and pass it with default parameters. .. GENERATED FROM PYTHON SOURCE LINES 21-29 .. code-block:: Python from scipy.stats import dirichlet from pycalib.visualisations.ternary import draw_func_contours function = lambda x: dirichlet.pdf(x, alpha=[5, 3, 2]) fig = draw_func_contours(function) .. image-sg:: /examples/images/sphx_glr_xmpl_ternary_contours_001.png :alt: xmpl ternary contours :srcset: /examples/images/sphx_glr_xmpl_ternary_contours_001.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 30-33 Next we show how do use a ternary calibration model that has 3 probability inputs and 3 ouputs. We will first simulate a calibrator by simulating 3 Dirichlet distributions and applying Bayes rule with equal prior. .. GENERATED FROM PYTHON SOURCE LINES 33-45 .. code-block:: Python class calibrator(): def predict_proba(self, x): pred1 = dirichlet.pdf(x, alpha=[3, 1, 1]) pred2 = dirichlet.pdf(x, alpha=[6, 7, 5]) pred3 = dirichlet.pdf(x, alpha=[3, 4, 5]) pred = np.vstack([pred1, pred2, pred3]).T pred = pred / pred.sum(axis=1)[:, None] return pred cal = calibrator() .. GENERATED FROM PYTHON SOURCE LINES 46-49 Then we will first draw a contourmap only for the first class. We do that by creating a lambda function and selecting the first column. We also select a colormap for the first class. .. GENERATED FROM PYTHON SOURCE LINES 49-54 .. code-block:: Python function = lambda x: cal.predict_proba(x.reshape(-1, 1))[0][0] fig = draw_func_contours(function, cmap='Reds') .. image-sg:: /examples/images/sphx_glr_xmpl_ternary_contours_002.png :alt: xmpl ternary contours :srcset: /examples/images/sphx_glr_xmpl_ternary_contours_002.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 55-58 We can look at the second class by creating a new lambda function and selecting the second column. We will also modify how many times to subdivide the simplex (subdiv=3). And the number of contour values (nlevels=10). .. GENERATED FROM PYTHON SOURCE LINES 58-62 .. code-block:: Python function = lambda x: cal.predict_proba(x.reshape(-1, 1))[0][1] fig = draw_func_contours(function, nlevels=10, subdiv=3, cmap='Oranges') .. image-sg:: /examples/images/sphx_glr_xmpl_ternary_contours_003.png :alt: xmpl ternary contours :srcset: /examples/images/sphx_glr_xmpl_ternary_contours_003.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 63-65 Finally we show the 3rd class with other sets of parameters and specifying the names of each class. .. GENERATED FROM PYTHON SOURCE LINES 65-71 .. code-block:: Python function = lambda x: cal.predict_proba(x.reshape(-1, 1))[0][2] fig = draw_func_contours(function, nlevels=10, subdiv=5, cmap='Blues', labels=['strawberry', 'orange', 'smurf']) .. image-sg:: /examples/images/sphx_glr_xmpl_ternary_contours_004.png :alt: xmpl ternary contours :srcset: /examples/images/sphx_glr_xmpl_ternary_contours_004.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 72-75 In order to plot the contours of all classes in the same figure it is necessary to loop over all subplots. We show an example that uses the previous functions. .. GENERATED FROM PYTHON SOURCE LINES 75-85 .. code-block:: Python labels=['strawberry', 'orange', 'smurf'] cmap_list = ['Reds', 'Oranges', 'Blues'] fig = plt.figure(figsize=(10, 5)) for c in [0, 1, 2]: ax = fig.add_subplot(1, 3, c+1) ax.set_title('{}\n$(C_{})$'.format(labels[c], c+1), loc='left') function = lambda x: cal.predict_proba(x.reshape(-1, 1))[0][c] fig = draw_func_contours(function, nlevels=30, subdiv=5, cmap=cmap_list[c], ax=ax, fig=fig) .. image-sg:: /examples/images/sphx_glr_xmpl_ternary_contours_005.png :alt: strawberry $(C_1)$, orange $(C_2)$, smurf $(C_3)$ :srcset: /examples/images/sphx_glr_xmpl_ternary_contours_005.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-timing **Total running time of the script:** (0 minutes 1.537 seconds) .. _sphx_glr_download_examples_xmpl_ternary_contours.py: .. only:: html .. container:: sphx-glr-footer sphx-glr-footer-example .. container:: binder-badge .. image:: images/binder_badge_logo.svg :target: https://mybinder.org/v2/gh/perellonieto/pycalib/gh-pages?filepath=notebooks/examples/xmpl_ternary_contours.ipynb :alt: Launch binder :width: 150 px .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: xmpl_ternary_contours.ipynb ` .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: xmpl_ternary_contours.py ` .. only:: html .. rst-class:: sphx-glr-signature `Gallery generated by Sphinx-Gallery `_