Finding the conditional distribution

Click For Summary
SUMMARY

The discussion focuses on deriving the conditional distribution of the variable ##u## given ##Y##, where ##Y## follows a Negative Binomial distribution and ##u## follows a Gamma distribution. The user has formulated the joint distribution and is attempting to simplify it to find ##f(u|y)##. Key equations provided include the joint probability and the use of parameters ##\alpha## and ##\beta##, defined as ##\beta = \phi## and ##\alpha = \phi^{-1}##. The user seeks guidance on the next steps to identify the resulting distribution after simplification.

PREREQUISITES
  • Understanding of Poisson distribution and its properties
  • Knowledge of Gamma distribution and its parameters
  • Familiarity with Negative Binomial distribution and its applications
  • Proficiency in Bayesian inference and conditional probability
NEXT STEPS
  • Research the derivation of conditional distributions in Bayesian statistics
  • Study the relationship between Negative Binomial and Poisson distributions
  • Explore the properties of Gamma distributions in the context of Bayesian inference
  • Learn about the use of moment-generating functions in finding distributions
USEFUL FOR

Statisticians, data scientists, and researchers involved in probabilistic modeling and Bayesian analysis will benefit from this discussion, particularly those working with conditional distributions and related statistical concepts.

Bazzinga
Messages
45
Reaction score
0
Hey guys, I'm trying to find a conditional distribution based on the following information:

##Y|u Poisson(u \lambda)##, where ##u~Gamma( \phi)## and ##Y~NegBinomial(\frac{\lambda \phi}{1+ \lambda \phi}, \phi^{-1})##

I want to find the conditional distribution ##u|Y##

Here's what I've got so far:

##f(u|y)= \frac{f(u, y)}{p(y)} = \frac{p(y|u)}{p(y)} f(u)##
##=\frac{(u \lambda)^{y}e^{-u \lambda}}{y!} \frac{u^{ \alpha - 1}e^{-u/ \beta}}{ \Gamma ( \alpha) \beta^{ \alpha}} \frac{y! \Gamma ( \alpha)}{\Gamma (y+ \alpha)} ( \frac{1+\lambda \beta}{ \lambda \beta})^{y} (1+ \lambda \beta)^{ \alpha}##

where ##\beta = \phi## and ## \alpha = \phi^{-1}##
(I'm using ##\beta## and ##\alpha## right now because that's what it does in the notes)

I'm not sure where to go from here. I can cancel some terms out, but then I'm not sure what distribution I'm supposed to end up with. Could anyone give me a push in the right direction?
 
Bazzinga said:
I can cancel some terms out
Then please do so and post what you get.
 

Similar threads

  • · Replies 10 ·
Replies
10
Views
2K
  • · Replies 5 ·
Replies
5
Views
1K
Replies
8
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
Replies
2
Views
2K
Replies
6
Views
3K
Replies
4
Views
2K
  • · Replies 26 ·
Replies
26
Views
3K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K