Add smoothing tutorial

[aleblanc30]

Hi !

I propose this Nadaraya-Watson/kernel smoothing tutorial, inspired from the kernel interpolation tutorial. It features:

• univariate regression

• univariate variance estimation

• bivariate regression on binary target

with a gaussian or a quadratic kernel.

[joanglaunes]

Hello @aleblanc30 ,
Thanks you very much, this is a very nice addition to the tutorials! I see that the tests didn't pass but I think it is completely unrelated. However we will need to fix this problem with the tests before we can merge.

[aleblanc30]

Thank you for the kind words ! Let me know if there is anything I need to do to help move the PR along 😄

[aleblanc30]

From what I can see it seems all the fee usage for Cirrus CI ran out, which killed the CI. Any chance someone can relaunch the CI ?

[bcharlier]

Hi @aleblanc30 . There is also a small problem with the CI pipeline in MacOs : conjunction of python3.12 and sequoia breaks the unit tests for a some reasons... we are on it.

Thank again for your contribution.

[aleblanc30]

I'm sorry I got impatient ! ^^

[joanglaunes]

Hello @aleblanc30 ,
I just merged some recent fixes into your PR, it should make the CI work. Now about your tutorial I have a few remarks/corrections to suggest :

• on the second figure, the legend says "Smoothed function" for the two dotted lines, could you correct this ?
• there are two kernel_smoother functions and two plot_smoothing_results functions; maybe having a single kernel_smoother which works for any dimension, and rename the others to plot_smoothing_results_1D and plot_smoothing_results_2D
• maybe changing a little bit the 2D example function, so that it is not exactly the same as the other tutorial.
  Would you be ok with these changes ?