1
$\begingroup$

I'm trying to generate a Weibull distribution to see if it fits some data. I have an arithmetic mean and variance for this data but I'm struggling to analytically create the specific Weibull parameters from these.

For a Weibull distribution:

μ=λΓ(1+(1/k))

σ^2=λ^2 [Γ(1+2/k)-(Γ(1+1/k))^2 ]

I have found that these equations rearrange to:

(σ^2/μ^2)+1=(2/k)!/((1/k)!)^2

This can be solved by Wolfram Alpha and with values of μ=7.3151 and σ=4.0730 this gives k=1.8648 and λ=8.2383.

The problem that I'm facing is that I don't understand the jump made by Wolfram from the rearranged equation to the answer. If anyone could offer an explanation that would be great.

$\endgroup$
  • $\begingroup$ You cannot obtain a unique solution for two variables from the single last equation. Generically, though, two equations in two unknowns have a discrete number of solutions, typically just one. Could you therefore expand on what you are looking for in way of an "explanation"? $\endgroup$ – whuber Apr 11 '17 at 16:31
  • $\begingroup$ Wolfram gives me 3 solutions to this equation. Two negative and one positive. The positive is the only useful solution. By "explanation" I mean that I would like to know how those answers are obtained. Does Wolfram calculate these iteratively or is there an analytical way to solve these? $\endgroup$ – Shardy Apr 11 '17 at 18:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.