|
So I am supposed to prove that $P(a)$ given $P(b \text{ and } a)$ is $1$ One way of solving this is that with $P(b \text{ and } a )$ as given the sample space has been reduced to that only, and for that sample space the $a$ is there always so answer is $1$. But when I tried to do that using bayes theorem, I couldn't solve it $P(b \text{ and } a \mid a)\cdot P(a)$ divided by $P(b \text{ and } a)$ Can anyone give me the explanation for this? |
||||
|
|
|
Bayes formula amounts to $p(x|y)p(y)=p(xy)$ by interchanging the roles of $x$ and $y$. So, here is a formal proof: $p(a|ab)\cdot p(ab)=p(aab)=p(ab)$. If $p(ab)\ne 0$ then you can divide by it to obtain $p(a|ab)=1$. |
|||
|
|
|
You simply need to come back to the definition of conditional probabilities. Given two events $A,B$, the probability of $A$ given $B$ is defined as $$P(A|B)=\frac{P(A\cap B)}{P(B)}.$$ Of course, as you have certainly seen, it corresponds to the intuitive notion of "conditonal probability": we restrict the sample space to $B$.
Now, apply this to compute $P(A | A\cap B)$. Bayes Therorem follows from a simple application of this definition. You usually use it when you want to compute $P(A|B)$ knowing $P(B|A)$, $P(A)$ and $P(B)$. |
|||
|
|
