comp.lang.idl-pvwave archive: archive » Fitting a Gaussian to an "unleveled" histogram: Interpreting the width

Home » Public Forums » archive » Fitting a Gaussian to an "unleveled" histogram: Interpreting the width

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: Fitting a Gaussian to an "unleveled" histogram: Interpreting the width [message #90255 is a reply to message #90218]

Sat, 14 February 2015 15:06

Craig Markwardt
Messages: 1869
Registered: November 1996

Senior Member

On Wednesday, February 11, 2015 at 7:52:09 PM UTC-5, Maryam wrote:
> Hello.
>
> I am trying to fit an inverted Gaussian function to a histogram that is higher at one end than the other. I simply define a function that is a Gaussian with negative amplitude (plus a constant to shift the Gaussian up along the y-axis) and call that function using mpfitfun to get the fit. It all works out fine and I do get a fit that looks reasonable. However, I am not sure how to interpret the standard deviation returned by mpfitfun.

It really depends why the histogram is "higher at one end than the other." Is it because there is a sloping continuum or background level? If so, why not fit a constant plus linear term to the background instead of just a constant? Make your model more realistic and you will get more realistic parameter values. Make your model more synthetic, and your parameters will be meaningless.

CM

Report message to a moderator

[Message index]

		Fitting a Gaussian to an "unleveled" histogram: Interpreting the width By: Maryam on Wed, 11 February 2015 16:52
		Re: Fitting a Gaussian to an "unleveled" histogram: Interpreting the width By: Craig Markwardt on Sat, 14 February 2015 15:06
		Re: Fitting a Gaussian to an "unleveled" histogram: Interpreting the width By: Maryam on Fri, 27 February 2015 12:25

Previous Topic:	Re: Function Graphics overlaid objects on image()
Next Topic:	Principle Componets Analysis

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Fri Oct 10 19:21:32 PDT 2025

Total time taken to generate the page: 1.27968 seconds