My discrete math text says, if $E_1$ and $E_2$ are independent events, and $E_1 | S$ and $E_2 | S$ are independent events, then $\text{P}(E_1 \cap E_2 | S) = \text{P}(E_1 | S) \cdot \text{P}(E_2 | S)$. But, what's the proof?

  • 1
    $\begingroup$ if $E_1$ and $E_2$ are...? $\endgroup$ – TheRealFakeNews Feb 5 '13 at 7:25
  • $\begingroup$ @AlanH independent events. Edited. Thanks. $\endgroup$ – f.nasim Feb 5 '13 at 7:27
  • $\begingroup$ Do you understand what independent mean ? $\endgroup$ – Damien L Feb 5 '13 at 7:35
  • 1
    $\begingroup$ What is the definition of "$E | S$"? Your text really says this is an event?? $\endgroup$ – Did Feb 5 '13 at 7:50
  • 1
    $\begingroup$ yes, and giving the correct statement of the problem is also important, if you expect an answer. $\endgroup$ – UwF Feb 5 '13 at 9:13

Conditional probabilities are defined by

$P(A|B)=\frac{P(A\cap B)}{P(B)}$ (where you assume $P(B)>0$).

Two events $A,B$ are called independent, if $P(A\cap B)=P(A)P(B)$.

If your text says that $E_1$ and $E_2$ are independent conditionally on $S$, then that simply means that they are independent w.r.t. the probability $P(\cdot|S)$, i.e. the relation that you want to prove,

$P(E_1\cap E_2|S)=P(E_1|S)P(E_2|S)$.

If your text says that all three events $E_1,E_2,S$ are independent (and $P(S)>0$), then you have

$P(E_1\cap E_2|S) =\frac{P(E_1\cap E_2\cap S)}{P(S)}=\frac{P(E_1)P(E_2)P(S)}{P(S)}=\frac{P(E_1)P(S)}{P(S)}\frac{P(E_2)P(S)}{P(S)}=\frac{P(E_1\cap S)}{P(S)}\frac{P(E_2\cap S)}{P(S)}=P(E_1|S)P(E_2|S)$

Please go back and read what exactly your text says. Then it should be easy to figure out the answer.

| cite | improve this answer | |
  • $\begingroup$ The problem statement clearly reads "...Assuming that $E_1$ and $E_2$ are independent events and that $E_1 | S$ and $E_2 | S$ are independent events,...". But, in the solution to the problem it reads "...Because of the assumed independence of $E_1$, $E_2$, and $S$, we have $P(E_1 \cap E_2 | S) = P(E_1 | S) \cdot P(E_2 | S)$,...". Now, it is clear $E_1$, $E_2$, and $S$ are independent actually; the text never said $E_1$ and $E_2$ are independent conditionally on $S$. So your second solution is the correct answer. Thanks. $\endgroup$ – f.nasim Feb 5 '13 at 14:12
  • $\begingroup$ I doubt very much that anybody would use the mysterious statement in your question to mean that $(E_1,E_2,S)$ are independent. $\endgroup$ – Did Feb 5 '13 at 15:27

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.