Search code examples
pythonscipycurve-fittingdata-fitting

SciPy fitting gives an error

I am fitting my function with experimental data. The function is complicated enough that I am unable to post here, but my fitting module looks like this:

out_put = scipy.optimize.leastsq(func, initial parameter, full_output=True, ftol=0.001, xtol=0.001, gtol = 0.001)
fitter_sol = out_put[0]
error = np.sqrt(out_put[1].diagonal())

The last line of code gives an error under execution, and the error looks like:

AttributeError: 'NoneType' object has no attribute 'diagonal'

What could be the potential source of this error?

Solution

  • The docs say the second result of leastsq is:

    None if a singular matrix encountered (indicates very flat curvature in some direction).

    So your input is a singular matrix.