What is the main Naive Bayes assumption?
Knowledge about the value of the class attribute indicates value of another attribute
Knowledge about the value of a particular attribute doesn’t tell us anything about the value of another attribute
Knowledge about the value of a particular attribute tells us everything we need to know about the class value
Even when the attribute independence assumption is violated, Naive Bayes scheme still works well.
In Naive Bayes – a prior probability is
Probability of event after evidence has been seen
Probability of an event before evidence has been seen
Naïve Bayes assumption allows for the evidence to be split into independent parts
Naive Bayes method would have a hard time learning from dataset that has
Many nominal attributes
Many numeric attributes
Many redundant attributes
Many irrelevant features