What is posterior predictive distribution?

Question: In Bayesian machine learning, what does the posterior predictive distribution represent? A) The maximum-likelihood estimate only B) The training loss after convergence C) The prior distribution before seeing data D) The distribution of future observations after integrating over posterior uncertainty Correct: D Explanation: The posterior predictive distribution accounts for uncertainty in model parameters by averaging predictions over the posterior distribution. It is useful when uncertainty estimates matter, not just point predictions. Topic: advanced ML / Bayesian learning / uncertainty