0
$\begingroup$

Let $X$ be a discrete random variable distributed according to $p_x(\cdot)$. Let $Y_1, Y_2, \ldots Y_N$ be discrete random variables that depend on $X$. Suppose that $Y_1, Y_2, \ldots, Y_N$ are independent and identically distributed conditioned on $X$, such as $$p_{Y_n\mid X}(y\mid x) = P_{Y\mid X}(y\mid x)$$ for $n = 1,2, \ldots, N$.

How could I apply Baye's rule to calculate $p_{X\mid Y_1, \ldots, Y_N}(x\mid y_1, \ldots, y_N)$ in terms of the given values $p_X(\cdot)$ and $p_{Y\mid X}(\cdot\mid\cdot)$?

I have been struggling with this application of Baye's rule and I would really appreciate any help.

$\endgroup$
0
$\begingroup$

I'm going to abuse notation by eliminating subscripts on $P$; that shouldn't cause any confusion. We have $$P(x|y_1,\cdots,y_n) = P(x , y_1 ,\cdots , y_n) / P(y_1 , \cdots , y_n).$$

As the $Y_i$'s are independent, the denominator is just the product of the Bayesian priors $P(y_i)$. Again, by independence of the $Y_i$'s, the numerator is $P(x) P(y_1|x)\cdots P(y_n|x)$. So you can write $P(x|y_1,\ldots,y_n)$ in the suggestive form $$P_X(x) \prod_{i=1}^n \frac{P_{Y|X}(y_i|x)}{P_Y(y_i)}.$$ In particular, you seem to need a prior distribution on $Y$ as well as on $X$.

$\endgroup$

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.