My Math Forum Maximum Likelihood estimation

 April 10th, 2017, 01:32 PM #1 Senior Member   Joined: Feb 2015 From: london Posts: 121 Thanks: 0 Maximum Likelihood estimation Trying to find MLE for $\displaystyle \lambda$ of the truncated Poisson distribution tPoisson(λ, k) when k=1 $\displaystyle p_x = \frac{\lambda^x e^{-\lambda}}{q_kx!}$ where: $\displaystyle q_k = 1- \sum_{i=0}^{k} \frac{\lambda^ie^{-\lambda}}{i!}$ $\displaystyle x = k+1, k+2, k+3......$ $\displaystyle \lambda > 0$ $\displaystyle i >= 2$ Log likelihood function: $\displaystyle ln(L(\lambda) = ln [\prod_{}^{\infty}\frac{\lambda^x e^{-\lambda}}{q_kx!} ]$ $\displaystyle ln(L(\lambda) = ln (\lambda) \sum_{}^{\infty} x - n\lambda - nln(q_k) - \sum_{}^{\infty} x!$ when k=1, $\displaystyle q_k = - \lambda e^{-\lambda}$ Normally I would differentiate $\displaystyle ln(L(\lambda)$ with respect to $\displaystyle \lambda$ and equate to 0, to find solution for $\displaystyle \lambda=$, however $\displaystyle ln(q_k)$ results in an error as you cant take natural log of negative numbers, so im not sure how to continue
 April 12th, 2017, 09:37 AM #2 Senior Member   Joined: Feb 2015 From: london Posts: 121 Thanks: 0 Please ignore this question, I went wrong further up the question stack, so dont need to do this.

 Tags estimation, likelihood, maximum

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post Actuary Calculus 0 July 13th, 2015 03:39 AM graphist_groupist Advanced Statistics 1 July 16th, 2012 10:28 PM 450081592 Advanced Statistics 0 February 21st, 2012 05:30 PM 450081592 Advanced Statistics 0 February 21st, 2012 04:50 PM Darkprince Advanced Statistics 0 October 27th, 2011 11:26 AM

 Contact - Home - Forums - Cryptocurrency Forum - Top