Treatment of variance [STAT] cubes

[Simon Conseil]

From @Azeret on September 4, 2018 17:52

It seems the ZAP does not apply to variance [STAT] cubes, only to the data ones.
Is there a way to change this?
After running the ZAP it results in original (therefore very large) variance [STAT] cubes that prevents usage and further analysis.

Thank you for response!
Tereza (tjerabko(at)eso.org)

Copied from original issue: ktsoto#4

[Simon Conseil]

@Azeret - Fyi, I moved the issue to the new repository, the other one is inactive.

There is no easy way to propagate the variance in the PCA process, so indeed it is not modified. But the propagated variance has anyway a number of issues, as described in http://adsabs.harvard.edu/abs/2017A%26A...608A...1B, so many people are either re-estimating or rescaling this variance. You can find our current solution in the paper above, but there are other slightly different approaches for the Muse Wide survey (http://adsabs.harvard.edu/abs/2017A%26A...606A..12H), and I think in http://adsabs.harvard.edu/abs/2016ApJ...831...39B, and probably others.

[Tereza Jerabkova]

Dear Simon, Thank you very much for very fast and very useful answer. Best wishes, Tereza - - - - - - - - - - - - - - - - - - - - - - - - - - - - Tereza Jerabkova, PhD student ESO – European Southern Observatory Karl-Schwarzschild-Straße 2, 85748 Garching bei München, Germany Phone [phone]‬ Mobile ‭[phone]‬ http://sirrah.troja.mff.cuni.cz/~tereza/, http://www.eso.org

…

On 5 Sep 2018, at 08:44, Simon Conseil ***@***.***> wrote: @Azeret <https://github.com/Azeret> - Fyi, I moved the issue to the new repository, the other one is inactive. There is no easy way to propagate the variance in the PCA process, so indeed it is not modified. But the propagated variance has anyway a number of issues, as described in http://adsabs.harvard.edu/abs/2017A%26A...608A...1B <http://adsabs.harvard.edu/abs/2017A%26A...608A...1B>, so many people are either re-estimating or rescaling this variance. You can find our current solution in the paper above, but there are other slightly different approaches for the Muse Wide survey (http://adsabs.harvard.edu/abs/2017A%26A...606A..12H <http://adsabs.harvard.edu/abs/2017A%26A...606A..12H>), and I think in http://adsabs.harvard.edu/abs/2016ApJ...831...39B <http://adsabs.harvard.edu/abs/2016ApJ...831...39B>, and probably others. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#5 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AVGusi3wkmPpzPgUydgSueDkSOUQJXTrks5uX3LSgaJpZM4WaRnF>.

[Simon Conseil]

You're welcome!