[SciPy-user] Questions on leastsq

[Stuart Wilkins]

Hi, 

I am having problems with the least squares fitting. This may or may not be a bug.

I am trying to get an idea of the errors from fitting data. I am usualy fitting to Gaussian Lorentzian, Lor**2 etc. When I try to compute the errors from

sigma = sqrt(diag(cov_x))

where cov_x is the covariance matrix I get results which are wildly different than anything before. Including some code I had a long time ago in Matlab.

Does anyone have any idea if the form of cov_x is correct or any success in using this?

Thanks,
Stuart