How does Naïve Bayes behave under Bayesian decision boundary instability in streaming data?

Naïve Bayes decision boundaries can drift in streaming environments due to incremental updates and distribution shifts.

In streaming data, Naïve Bayes continuously updates counts and probabilities. If data distribution changes, posterior decision boundaries shift dynamically. Without drift correction or forgetting mechanisms, this leads to unstable predictions and degraded long-term performance.