These are chat archives for astropy/astropy

1st
Nov 2017
Thomas Spriggs
@tspriggs
Nov 01 2017 11:52
Hi all, hoping someone can help me with a query about the fitting side of Astropy.... I am fitting emission spectra with a 1D gaussian model, I know where the peak should be in the data and hence give it as an initial guess, but when I run it through fitter, it doesn't seem to want to vary to much side to side about where the final Mean (wavelength) should be for the actual peak.... in a custom model I have set limits on where it can look, but it will still not take in to consideration 5003, for example, as where the peak can be, it should be at 5007 but due to shifts in wavelength it is nudged a bit... My query is how can I understand why it's not looking further afield? and how could I make it do so? I am using the LevMar fitter by the way
matrixbot
@matrixbot
Nov 01 2017 15:17
astrofrog (@_slack_astropy_U07529085:matrix.org) What is the flux level of the spectrum? (peak value)
astrofrog (@_slack_astropy_U07529085:matrix.org) I've seen issues in the past when the peak is e.g. 1e-13 since some of the fitters work with an absolute tolerance
astrofrog (@_slack_astropy_U07529085:matrix.org) See astropy/astropy#6269 specifically
astrofrog (@_slack_astropy_U07529085:matrix.org) By the way, we've moved from Gitter to Slack, so feel free to join slack using http://joinslack.astropy.org and then accessing slack itself using http://astropy.slack.com