Please enable JavaScript.
Coggle requires JavaScript to display documents.
Naive Bayes (Naive Bayes (Kind of classifier which
uses the Bayes Theorem…
Naive Bayes
Conditional Probability
-
P(A|B)
- Probability of A under the condition B
Bayes Theorem
-
- P(H) is prob. of hypothesis H being true / Prior Probability
- P(E) is prob. of evidence
Bayesian Classifier
-
-
-
Since P(X) is constant for all classes, only needs to be maximized
Naive Bayes
-
Maximum A Posteriori (MAP)
- Highest Probability is considered as
the most likely class
- MAP(H)
= max(P(H|E))
= max(P(E|H)*P(H))
-
-
-
Type
Gaussian
- Used in classification & it assumes
that features follow a normal distribution
-
MultiNomial
- use on data that is
multinomially distributed
-
-
Bayesian Belief Network
-
Graphical model of causal relationships
- Node = Random variable
- Links = Dependency
- Has no loops/cycle
Joint Probability
- P(G,S,R) = P(G|S,R)P(S|R)P(R)
- Assumption
Parameters of problem are only dependent on the
probabilities of their parents.
- Called conditional probability