Request: add Knuth and Bayesian blocks binning methods to histogram

The astropy package hast two very useful bin methods that could be added to histogram:

• knuth_bin_width
• bayesian_blocks

[bsipocz]

There is clear consensus above that these don’t belong to numpy ,so I’m going to close this issue. If there is a notion that these would be a net benefit to scipy (most bugs got already smoked out of them while the code was in astroML and then in astropy), and if there is a want for it, then I’m happy to do the upstreaming myself as I have some familiarity with the code itself.

[madphysicist]

I think this would be fine for scipy, especially since the building blocks are already there, but I will defer to people with actual experience in the matter.

While we are all gathered here, I would like to ask if there is any interest in Wand’s estimator. It is supposedly as good as you can get in terms of MISE: http://www.stat.cmu.edu/~rnugent/PCMI2016/papers/WandBinWidth.pdf, found via http://stats.stackexchange.com/a/55205/102011. The implementation I am thinking of would have an additional order parameter that would have a default value.