this post was submitted on 25 Nov 2023
1 points (100.0% liked)
Machine Learning
1 readers
1 users here now
Community Rules:
- Be nice. No offensive behavior, insults or attacks: we encourage a diverse community in which members feel safe and have a voice.
- Make your post clear and comprehensive: posts that lack insight or effort will be removed. (ex: questions which are easily googled)
- Beginner or career related questions go elsewhere. This community is focused in discussion of research and new projects that advance the state-of-the-art.
- Limit self-promotion. Comments and posts should be first and foremost about topics of interest to ML observers and practitioners. Limited self-promotion is tolerated, but the sub is not here as merely a source for free advertisement. Such posts will be removed at the discretion of the mods.
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
The interesting challenge is trying to figure out how you solved the problem to get 1/4 instead of 1/2. In Bayesian thinking, you have the prior and posterior. The prior (before you see the evidence that a = 1, b =1, and c = 0) is the K column by itself. P(K = 1) is 1/2 as there are 4 ones and 4 zeros.
Now the posterior is evaluated with respect to the prior. In Naive bayes, it is the case that the pieces of evidence are viewed independently (naively) from each other. So P(K = 1 | a = 1 and b = 1 and c = 0) is simplified as P(K = 1) P(a = 1 and b = 1 and c = 0 | K = 1) / P(a = 1 and b = 1 and c = 0). The numerator simplifies to 1/2 * P(a = 1 | K = 1) * P(b = 1 | K = 1) * P(c = 0 | K = 1) = 1/2 * 1/2 * 1/4 * 1/2.
The denominator is again, challenging. If you calculate it like you should (not-naively), it should equal P(a = 1 and b = 1 and c = 0). But the problem becomes that... that will make all the probabilities over the conditional distribution sum up to 1 if you are to calculate it non-naively (i.e., *not* assuming P(a = 1 and b = 1 and c = 0 | K = 1) = P(a = 1| K = 1) * P(b = 1 | K = 1) * P(C = 0 | K = 1)).
The way the solution is calculating it sidesteps this issue by expressing P(a = 1 and b = 1 and c = 0) in such a way that is amenable to Naive Bayes. Think about this further.