1
$\begingroup$

Medical investigations show that symptoms $S_1$ and $S_2$ can occur at exactly $3$ illnesses $I_1, I_2$ and $I_3$, with (conditional) probabilities $m_{i,j} = P(S_j \mid I_i)$ with $i=1,2,3$ and $j=1,2$. In matrix form they look like this: $M = (m_{ij})=\begin{pmatrix} 0.8 & 0.3\\ 0.2 & 0.9\\ 0.4 & 0.6 \end{pmatrix}$

The prior probabilities for illnesses $I_1, I_2, I_3$ are given by the vector $(0.3, 0.6, 0.1)$. Assume the illnesses exclude each other.

Determine the conditional probabilities $P(I_i \mid S_j)$, $j=1,2$, $i=1,2,3$


I think better first change $$P(I_i \mid S_j)=\frac{P(I_i \cap S_j)}{P(S_j)}= \frac{P(I_i)}{P(S_j)}$$

Now I need calculate it for every index using this formula right?

For example for $i=j=1: \frac{P(I_1)}{P(S_1)}= \frac{0.3}{0.8}= 0.375$

Next we do for $i=1, j=2: \frac{P(I_1)}{P(S_2)}= \frac{0.3}{0.3}= 1$

etc.

But not sure if this is good?

$\endgroup$
1
$\begingroup$

Hint: Use Bayes' Rule.


$$P (I_i \mid S_j) = \frac {P (I_i)P (S_j \mid I_i)}{\sum_{i=1}^{3}P (I_i)P (S_j \mid I_i)}$$

$\endgroup$

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.