How does Naïve Bayes perform probability calibration and why is it often poorly calibrated?

Naïve Bayes often produces poorly calibrated probabilities due to independence assumptions.

Probability calibration refers to how well predicted probabilities reflect true likelihoods. Naïve Bayes tends to be overconfident because multiplying independent probabilities exaggerates certainty. Techniques like Platt scaling or isotonic regression can improve calibration.