midNaïve Bayes
What is the role of likelihood in Naïve Bayes classification?
Updated May 17, 2026
Short answer
Likelihood measures how probable observed features are given a class.
Deep explanation
Likelihood P(X|C) quantifies how likely features X are generated under class C. Naïve Bayes assumes conditional independence so likelihood becomes product of individual feature probabilities. This simplifies high-dimensional probability estimation.
Real-world example
Words like 'win', 'score' increase likelihood of sports category.
Common mistakes
- Confusing likelihood with posterior probability.
Follow-up questions
- Why multiply probabilities?
- What is log-likelihood?