GLM

[Devanshi Maheshwari]

Y1, Y2....... Yn are independent claims, which are assumed to be exponentially distributed, with E(Y)- μι
i) Show that the canonical link function is the inverse link function.
(ii) It is decided that the canonical link should not be used, but that the mean claim sizes should be modeled as follows:
log mu i = alpha i=1,2,...,m beta i = m + 1 ,m+2,...,n
(a) . Derive the log-likelihood
(b)
Derive the max likelihood estimators of alpha and beta
(c) scaled deviance for this model
(iii) For a particular data set, m = 20 n = 44 1/20 sum l = 1 to 20 y l =14.2 Calculate the deviance for y_{1} = 7 1/24 sum i = 21 to 44 y i =18.7

 

[John Lee]

Hi there - this forum for helping students - not for tutors to do the work instead of the students!

So, show me what you've done so far and then I can help you find your errors in your work and help you understand why it's wrong.

 

[Devanshi Maheshwari]

Hi there - this forum for helping students - not for tutors to do the work instead of the students!

So, show me what you've done so far and then I can help you find your errors in your work and help you understand why it's wrong.

Hi I completed most of the sums but when it comes to scaled deviance it gets difficult

 

[Andrea Goude]

Screen shot or take a photo of your solutions and we can help you with the next steps.