0
$\begingroup$

Although this is a similar question, Factored Joint Distribution of Tree Augmented Naive Bayes Algorithm, I need additional clarification.

I have training data for two different classes, so it is fairly simple to calculate the priors for each attribute $A_i$, given the class $C$.

I am trying to determine the posterior probability from a Tree Augmented Naive Bayes algorithm.

$P(\mathrm{Class}|\mathrm{Attributes}) = P(\mathrm{Class})\cdot P(\mathrm{Root}|\mathrm{Class})\cdot \prod_i P(A_i | \mathrm{parent}, \mathrm{Class})$.

I have $P(C)$, $P(\mathrm{Root}|C)$ and all $P(A_i |C)$, however, how do I find $P(A_i |A_p, C)$ if $\mathrm{parent}$ is not equal to $\mathrm{root}$?

$\endgroup$
0
$\begingroup$

I think I may have answered my own question. Since I have the training data, I counted the occurrences for both $A_i$ and $A_\mathrm{parent}$ with the class $C = 0$ and $C = 1$.

I created then my conditional tables.

$\endgroup$

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.