1
$\begingroup$

I have a table of 3 binary variables whose joint probability is given.

a b c p(a,b,c)
0 0 0 0.192
0 0 1 0.144
0 1 0 0.048
0 1 1 0.216
1 0 0 0.192
1 0 1 0.064
1 1 0 0.048
1 1 1 0.096

I see that there is a joint probabilities formula for 3 variable

$P(a,b,c) = P(c|a,b)P(a,b) = P(c|a,b)P(b|a)P(a)$

But when I cross-validate it with the table value and the formula it's not exact. What's happening here I am not getting it. It's not the same formula I assume. But how do I get the conditional probabilities from this table? If I can get the form, I can build the DAG.

I can see some patterns like 0.192 and 0.048 is twice even though $a$ is changed. Then I thought maybe $a$ is independent of $b$ and $c$?

I also find this formula $p(θ|X,α)=\frac{p(X|θ)p(θ|α)}{p(X|α)}$

So, I tried with a=1, b=1, c=1,

$ P(c|a, b) * P(b|a) * P(a) = P(c|a, b) * \frac{P(b \cap a)}{P(a)} * P(a) = \frac{P(c\cap a \cap b)}{P(a\cap b)} * \frac{P(b \cap a)}{P(a)} * P(a) = \frac{1}{8} $

$\endgroup$
10
  • 1
    $\begingroup$ None of $a,b,c$ are independent of the others. Can you show use what you got when you tried $P(a,b,c) = P(c\mid a,b)P(a,b) = P(c\mid a,b)P(b\mid a)P(a)$ for some particular value of $a,b,c$? $\endgroup$
    – Henry
    Sep 29, 2022 at 18:26
  • 1
    $\begingroup$ The answer to the title question is "apply a definition of conditional probability." For instance, you can compute $P(b\mid a)=P(a\text{ and } b)/P(a)$ and read off the probabilities for any possible values of $(a,b)$ directly from the table. $\endgroup$
    – whuber
    Sep 29, 2022 at 19:25
  • $\begingroup$ @Henry I have added my approach. maybe it's the wrong approach. But the thing I am trying to do is to build that DAG from the formula or vice versa. I am trying to find out how that p(a, b, c) is generated. And also how did you find out none of a, b, or c is independent of each other? I also got the same observation. Just curious about your approach. $\endgroup$ Sep 29, 2022 at 20:38
  • $\begingroup$ That does not show me your calculations. I would say $\frac23 \times \frac9{25}\times \frac25 = \frac{12}{125}$ not $\frac18$ $\endgroup$
    – Henry
    Sep 30, 2022 at 1:06
  • $\begingroup$ @Henry How did you get these values? $\endgroup$ Sep 30, 2022 at 5:44

1 Answer 1

0
$\begingroup$

First, It appears that you are mixing up the definition of conditional probability for events and for random variables.

The definition $P(a,b,c)=P(c|a,b)P(a,b)=P(c|a,b)P(b|a)P(a)$ is for events.

For random variables, the definition is :

$$P(A = a,B = b,C =c)=P(C = c|A = a,B = b)P(B = b,A = a)$$

where $a,b,c$ is the values that the random variables $A$, $B$, $C$ can assume (in this case, $0$ or $1$).

Now, if you want the conditional probability of $C$, given $A$ and $B$, you must obtain the probability $P(B = b,A = a)$.

Here is what you should do:

First, obtain $P(B = b, A = a)$ for all possible values of $a$ and $b$. For example, to obtain $P(B = 0, A = 1)$, just sum all probabilities where $B = 0$ and $A = 1$. Do the same for the other possible values of $A$ and $B$.

  1. Now, you can calculate $P(C = c|A = a,B = b) = \frac{P(A = a,B = b,C =c)}{P(B = b,A = a)}$

you can follow the same reasoning for the other conditional probabilities.

A good exercise for learning is to obtain all possible combinations of conditional and unconditional probabilities for all variables.

$\endgroup$
2
  • 1
    $\begingroup$ The definitions for random variables and events are the same, it's just a matter of notation. People sometimes do use notation shortcuts like using $P(a)$ meaning $P(A=a)$. $\endgroup$
    – Tim
    Jun 8, 2023 at 5:51
  • $\begingroup$ @Tim. Yes, I understand and agree. Maybe you could help me to rewrite the first sentences in a more precise way, I would appreciate it. A side note: In my understanding, the OP was confusing about the events definition, for this reason I introduced the random variables. Another possible approach it would be use the notation $a_0$ and $a_1$ to represent, respectively, the events $\{w \in \Omega: a=0\}$ and $\{w \in \Omega: a=1\}$. But I think this could lead to more confusion. $\endgroup$ Jun 8, 2023 at 13:11

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.