A naive Bayes classifier assigns an observation with features to a class by applying Bayes' theorem together with the assumption that the features are conditionally independent given the class. Thus it chooses the class maximizing

The independence assumption is often false, but the resulting classifier can still perform well and is inexpensive to train. Discrete, Gaussian, and other models differ in how the conditional feature distributions are estimated.

See also

Bayes' Theorem, Bayesian Analysis, Conditional Probability

Explore with Wolfram|Alpha

References

Hastie, T.; Tibshirani, R.; and Friedman, J. The Elements of Statistical Learning, 2nd ed. New York: Springer-Verlag, 2009.

Cite this as:

Weisstein, Eric W. "Naive Bayes Classifier." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/NaiveBayesClassifier.html

Subject classifications