Solve $P(X_1 = k |X_1 + X_2 = n)$ poisson distribution

278 Views Asked by At

Find distribution of $X_1$ by solving $P(X_1 = k |X_1 + X_2 = n)$ where $X_1, X_2$ are independent Poisson variables with parameters $\lambda_1, \lambda_2$ and given $X_1 + X_2 = n$

Solution:

$$P(X_1 = k |X_1 + X_2 = n) = \frac{P(X_1 = k, X_1 + X_2 = n)}{P(X_1 + X_2 = n )}$$

$$= \frac{P(X_1 = k)P(X_2 = n - k)}{P(X_1 + X_2 = n)}$$

$$= \frac{\frac{\lambda_1^ke^{-\lambda_1}}{k!}\frac{\lambda_2^ke^{-\lambda_2}}{(n-k)!}}{e^{-(\lambda_1+\lambda_2)}(\lambda_1 + \lambda_2)^n/n!}$$

$$=\frac{n!}{k!(n-k!)}\left(\frac{\lambda_1}{\lambda_1+\lambda_2} \right)^k\left(\frac{\lambda_2}{\lambda_1 + \lambda_2} \right)^{n-k}$$

This is binomial theorem.


I don't understand the last step.

This part

$$\left(\frac{\lambda_1}{\lambda_1+\lambda_2} \right)^k\left(\frac{\lambda_2}{\lambda_1 + \lambda_2} \right)^{n-k}$$

What I get while trying this was

$$\left(\frac{\lambda_1^k}{(\lambda_1+\lambda_2)^n} \right)\left(\frac{\lambda_2^{n-k}}{(\lambda_1 + \lambda_2)^n} \right)$$

Which is what it equals to I'm not sure how the $n$ was replaced with $k$ and $(n-k)$.

Also on the second step

$$\frac{P(X_1 = k, X_1 + X_2 = n)}{P(X_1 + X_2 = n )} = \frac{P(X_1 = k)P(X_2 = n - k)}{P(X_1 + X_2 = n)}$$

Couldn't they cancell the denominator in the same way ? What property lets them split the AND part of P like $P(A, B) = P(A)P(B)$

1

There are 1 best solutions below

0
On BEST ANSWER

$(X_1=k,X_1+X_2=n)=(X_1=k, X_2=n-k)$ and the events $(X_1=k),( X_2=n-k)$ are independent. His justifies the second step.

In line three you have mis-typed $\lambda_2^{k}$ for $\lambda_2^{n-k}$.

In the denominator they have split $(\lambda_1+\lambda_2)^{n}$ as $ (\lambda_1+\lambda_2)^{k}(\lambda_1+\lambda_2)^{n-k}$.