How does Naïve Bayes interact with class imbalance in extreme skew distributions?

Naïve Bayes is sensitive to class imbalance because priors strongly influence posterior probabilities.

In extreme imbalance, P(C) dominates posterior computation. Rare classes may be overshadowed unless priors are adjusted or resampled. Techniques like prior reweighting, synthetic sampling, or threshold adjustment are used to mitigate bias. Despite this, NB is often more stable than many classifiers under imbalance.