0
$\begingroup$

$P$ and $S$ are the common cause of $c$. If $P(C=true| P,S )$ is given as the table below, and $P(S=true) =0.3$, $P(P=true) =0.9$ how can I eliminate $S$ and calculate $P(C=true | P=true )$ and $P(C= false| P=true)$ ? Many thanks for your help.

$P(C=true | P,S)$:

+p  +s +c  0.05
+p  -s +c  0.02
-p  +s +c  0.03
-p  -s +c  0.001  
$\endgroup$

1 Answer 1

1
$\begingroup$

You need to sum out a variable using the law of total probability. So $$ P(c=true | p=true) = P(c=true | p=true, s=true) * P(p=true, s=true) + P(c=true | p=true, s=false) * P(p=true, s=false) $$ This can intuitively be understood as summing probabilities across mutually exclusive spaces. This Stanford pdf explains where the equation comes from well.

Because p and s are independent

$$ P(p=true, s=true) = P(p=true) * P(s=true) $$ $$ P(p=true, s=false) = P(p=true) * P(s=false) $$ So do the substitution in the above equation and you're good to go.

$\endgroup$
6
  • $\begingroup$ I know I can sum out P(C, P,S) , can I sum out P(C | P, S) as well? How to calculate P(C=false | P=true) then? I am confused as if I use P(C,P,S) = P(P,S)*P(C|P,S) , and sum out P(C, P,S), the results are different from suming up P(C, P,S) directly. Which one is correct? Thanks! $\endgroup$ Commented May 26, 2019 at 3:58
  • $\begingroup$ Sorry, read your question too quickly and assumed your table was the joint distribution. Have updated the answer $\endgroup$
    – Brent
    Commented May 26, 2019 at 4:29
  • $\begingroup$ Thanks. could you explain a bit more why the first equation ? $\endgroup$ Commented May 26, 2019 at 8:31
  • $\begingroup$ should the last bit be P(s= false, p= true)? $\endgroup$ Commented May 26, 2019 at 8:32
  • $\begingroup$ I mean should it be +𝑃(𝑐=𝑡𝑟𝑢𝑒|𝑝=𝑡𝑟𝑢𝑒,𝑠=𝑓𝑎𝑙𝑠𝑒)∗𝑃(𝑠=𝑓𝑎𝑙𝑠𝑒,𝑝=true) ? $\endgroup$ Commented May 26, 2019 at 10:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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