- The Naive Bayes classifier is a probabilistic classifier that applies Bayes' theorem with strong (naive) independence assumptions.
- It estimates the probability of a class given attribute values by calculating the product of the individual attribute probabilities multiplied by the class prior probability.
- Training involves counting occurrences of attribute-value pairs in the training data to estimate probabilities, with smoothing to address zero counts. Classification then predicts the class with highest posterior probability for new examples.