How does Naïve Bayes interact with mutual information-based feature weighting?

Mutual information can be used to weight features in Naïve Bayes by measuring dependency strength with the class label.

Mutual information quantifies how much knowing a feature reduces uncertainty about the class. In weighted Naïve Bayes, each feature likelihood is exponentiated or scaled by its MI score, effectively emphasizing informative features while suppressing noisy ones. This partially relaxes the independence assumption by introducing feature importance weighting.