Hans Moritz Günther

Results 55 issues of Hans Moritz Günther

*This PR is not intended to be merged (at least not in the current form). It's purpose is to show a possible implementation of models that depend on more that...

long-term vision

In a chi^2 statistic, the goodness of fit can be derived analytically from the chi^2 value and the number of DOF. That's not true for many of our other statistics...

type:feature
general:statistics

[Arviz](https://python.arviz.org/en/latest/index.html) is a Python package for "Exploratory analysis of Bayesian models". It provides a lot of plotting and quicklook capability and seems to be a becoming a kind of standard...

type:enhancement
general:visualization

We can link the parameters of model, e.g. `g2.pos = g1.pos + 0.125`. We can also set the max and min of a parameter, e.g. `g2.pos.max = 11.234`. However, I...

type:enhancement
type:feature
general:optimisation

This is based on https://github.com/astropy/astropy/issues/14915 The build system is not necessarily my area of expertise, but if other, similar packages (in this case astropy) work out how to simplify things,...

area:build

This is a summary of discussion in #1650 which introduced a possible implementation, which we decided not to merge in this form. See #1650 for more details. ## Motivation Literally...

type:feature

Using `plot_fit` if might be useful to show data and model in the same color, e.g,. in the following plot, it's not clear if the red model is supposed to...

dep:matplotlib
general:visualization

The current plotting backend has: ```Python def set_subplot(self, row, col, nrows, ncols, clearaxes=True, left=None, right=None, bottom=None, top=None, wspace=0.3, hspace=0.4): ``` and ```Python def set_jointplot(self, row, col, nrows, ncols, create=True, top=0,...

type:question

This was noticed while working on #1382, but we decided there that it's out of scope for that PR, so I'm opening a separate issue. Our current pylab backend for...

general:visualization

I recently added the ability to change I/O and plotting backends. In both cases, I wired that into the existing Sherpa infrastructure by essentially monkey-patching `sherpa.io.backend` and `sherpa.plot.backend` when a...