2

When does naive Bayes perform poorly? Can you think of any specific examples of problems in which it wouldn't work? We can ignore not having seen given data points before as that can be corrected by Laplace smoothing.

usεr11852
  • 33,608
  • 2
  • 75
  • 117
Remertion
  • 21
  • 1
  • 2

1 Answers1

1

The answer lies in the naive Bayes conditional independence assumption: When features are not independent given the class label, naive Bayes will make wrong decisions.

For example, in spam classification, using naive Bayes can lead to poor performance (for details read about bayesian poisoning)

shyamupa
  • 383
  • 2
  • 9
  • In all fairness, the conditional independence assumption is violated in nearly all cases where naive bayes is used, and the technique often performs well anyway. That's why we call it "naive," because we know the independence assumption is generally not justified. – David Marx Nov 25 '13 at 22:28