0
$\begingroup$

I've already worked on a couple of random variables with their value in $\mathbb{N}$ $(X,Y)$

$p_{kn}=P(X=k,Y=n)= \left( \frac{λ^ke^{-y} α^n(1-α)^{k-n}}{n!(k-n)!} \right)\mathbb{1}_{\{0\le n\le k}\} (k,n)$

I had to find the law of X, of Y, trying to show that X and Y are independent...

Here is the point: the assumption is that the number of children for a random family follows a Poisson law with $E(Y)=2$. Assuming that for every birth, the probability to observe a boy's birth is equal to $1/2$. Show that, conditionally to the fact that the family has n children, the boys number follows a a Binomial law $B(n,1/2)$

I answered that conditionally to the fact that a family has $Y=n$ children, Y being the random variable representing the possible outcomes of the number of children inside a family, the boys number ,represented by X is a repetition of independent events identically distributed following a Bernouilli law with the probability $p=1/2$.

Thus $P(X=k|Y=n)~B(n,1/2)$ is the number of boys in a family conditonaly on the fact that it has n children.

But is that answer sufficient to prove P(X=k|Y=n)~B(n,1/2)?

$\endgroup$
0
$\begingroup$

Yes, that would be sufficient to argue that $X\mid Y=n \sim \mathcal {Bin}(n, \alpha)$.

However, the expression you provided is not quite right.   Where did the (lowercase) $y$ term come from?   Why are you raising $\lambda$ to the power of $k$?   You should have obtained:

$$\begin{align} p_{k,n;\alpha,\lambda} & = \mathsf P(X=k, Y=n; \alpha, \lambda) \\[1ex] & = \mathsf P(X=k \mid Y=n; \alpha)\cdot\mathsf P(Y=n; \lambda) \\[1ex] & = \binom{n}{k}\alpha^k(1-\alpha)^{n-k}\cdot\frac{\lambda^n \mathsf e^{-\lambda}}{n!}\;\mathbf 1_{0\leq k\leq n} \\[1ex] & = \frac{\alpha^k(1-\alpha)^{n-k}\lambda^n\mathsf e^{-\lambda}}{k!\,(n-k)!}\;\mathbf 1_{0\leq k\leq n} \end{align}$$

Where $\alpha = \tfrac 1 2$ is the success rate (of obtaining a boy each birth), and $\lambda=2$ is the average rate of births.


PS: You can now show that $X$ and $Y$ are not independent.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.