The logic behind HighDensityRegionAlgorithm.plotContour is weird.

[mbaudin47]

The logic behind HighDensityRegionAlgorithm.plotContour is weird.

This is the current code:

• plotContour(plotData=False, plotOutliers=False) : plots contours and outliers
• plotContour(plotData=False, plotOutliers=True) : plots contours and inliers
• plotContour(plotData=True, plotOutliers=True) : plots contours, outliers and inliers

The two functionnalities should be separated into three methods:

• drawContour() : plots only the contour
• drawInliers() : plots the inliers
• drawOutliers() : plots the outliers

[tupui]

OK to separate this method into three methods. @mbaudin47 Then what is the desired API for ProcessHighDensityRegionAlgorithm?

[tupui]

I have done a proposition in #9 .

[tupui]

Do pairplot in n-dimension using the 2D distribution using self.distribution.getMarginal([i, j])

[mbaudin47]

Pour créer un subplot contenant des graphiques OT, voir "Annexe : dessiner plusieurs distributions dans un même graphique" dans :

https://github.com/mbaudin47/otsupgalilee-eleve/blob/master/1-Intro-OT/Distributions.ipynb