0

If I need to compute $P(y)$ and I only know $P(y|x)$ and $P(x)$.

I know that:

$$P(y|x)=\frac{P(y,x)}{P(x)}$$

This is the contingency table for $P(y|x)$:

 P(y|x) |   +x   |  ¬x  |
-------------------------
   +y   |  0.91  | 0.03 |
-------------------------
   ¬y   |  0.09  | 0.97 |
-------------------------

And the probability of $x$:

        |   +x   |  ¬x  |
-------------------------
  P(x)  |  0.05  | 0.95 |
-------------------------

How can I compute $P(y)$?

  • There is insufficient information - for example if P(y|x) were 1, that could mean that P(y) =1 or it could also mean P(y) = P(x) if totally dependant. – Cato Jun 03 '19 at 15:27
  • @Cato, I have updated my question. If you need more info, please tell me. Thanks. – VansFannel Jun 03 '19 at 15:34
  • You have an equation $a=b/c$ and you know $a$ and $c$. So you can get $b$ just by multiplying both sides of the equation by $c$. – Michael Jun 03 '19 at 15:39

1 Answers1

1

$P(y,x) = P(y|x) P(x)$

$P(y) = \sum_x P(y,x) = \sum_x P(y|x) P(x) $

NicNic8
  • 6,951