What are the Pros/Cons of Naive Bayes? 

Pros:

  • Computational efficiency
  • Less prone to overfitting
  • If assumption of independence between features holds, the algorithm often has superior performance to other classification techniques
  • Highly suitable when all features are categorical, such as in text classification

Cons:

  • Independence assumption is not realistic for many data sets, and if that is the case, the algorithm suffers from high bias