Calculation of Jeffreys Prior for a Poisson Likelihood. These short videos work through mathematical details used in the Multivariate Statistical Modelling module at UWE.
My lecturer uploaded a powerpoint presentation 'explaining' Jeffrey's prior, plus complete lecture notes on Bayesian Statistics, plus a problem sheet solely of Jeffrey's prior. In under 20 minutes of video (this and the Binomial video) you've explained it far, far better than he has in all of that combined. Thank you for this!
Thank you very much! I was trying to do that using the other expression of the Information {sqrt[E[deriv(log(f(y|theta)))]^2]} but I couldn't go on! It's easier this way, now it works! Bye from Italy!
correct me if i am wrong that fisher information should be calculated for one copy of Y, so n should be 1 in the first place, instead of making n to be 1 in the last step.
Hello, good video! Do you know whats the family of the posterori Distribuion is when you use Jeffreys Prior as Prior? Its is Poisson again, because Jeffreys Prior isnt a informative prior?
Its a gamma distribution... as the gamma distribution in the conjugate prior of the poisson distribution [the Jeffrey's Prior can be considered as in improper form of the gamma distribution. The resulting posterior is a gamma with parameters r= (\sum y)+.5 and v = n
My lecturer uploaded a powerpoint presentation 'explaining' Jeffrey's prior, plus complete lecture notes on Bayesian Statistics, plus a problem sheet solely of Jeffrey's prior.
In under 20 minutes of video (this and the Binomial video) you've explained it far, far better than he has in all of that combined.
Thank you for this!
Thank you soooo much. clear derivation
thank you very much!
Thank you very much! I was trying to do that using the other expression of the Information {sqrt[E[deriv(log(f(y|theta)))]^2]} but I couldn't go on! It's easier this way, now it works!
Bye from Italy!
correct me if i am wrong that fisher information should be calculated for one copy of Y, so n should be 1 in the first place, instead of making n to be 1 in the last step.
Hello, good video! Do you know whats the family of the posterori Distribuion is when you use Jeffreys Prior as Prior?
Its is Poisson again, because Jeffreys Prior isnt a informative prior?
Its a gamma distribution... as the gamma distribution in the conjugate prior of the poisson distribution [the Jeffrey's Prior can be considered as in improper form of the gamma distribution. The resulting posterior is a gamma with parameters r= (\sum y)+.5 and v = n
deetoher Thank you very much ! Greetings from Germany!
Hi could I ask why is the expected value n*mu?
gül hocadan gelenler?