0

I know I can fit Weibull Distribution like this:

weibull_params = sp.stats.exponweib.fit(data, floc=0, f0=1)

But what method does it use? From Fitting a Weibull distribution using Scipy, it seems like using Max Likelyhood Method. Is it possible to switch to Least Square Method?

Addtionally, I want to know the quality of each fitting method(the error?), but I guess I can google it out and save it for later.

Community
  • 1
  • 1
ZK Zhao
  • 19,885
  • 47
  • 132
  • 206
  • *"...it seems like using Max Likelyhood Method."* That is correct, the `fit` method uses maximum likelihood estimation. – Warren Weckesser Aug 15 '15 at 03:01
  • Saullo Castro's answer in the link you provided shows a code example on how to switch to LSQ. – Dux Aug 17 '15 at 16:18
  • 1
    Is it possible, yes of course. But which way do you want to compute mean square distance? That is, do you want to fit the empirical distribution function of the data to Weibull CDF, or empirical density function to Weibull pdf, or empirical hazard function to Weibull hazard function? Each of those different methods will likely result in different result. And~, why not using MLE? – CT Zhu Aug 19 '15 at 20:19

0 Answers0