What is the infinite ensemble interpretation of Random Forest?

Random Forest can be viewed as averaging predictions over an infinite population of randomized trees.

The theoretical view treats Random Forest as an expectation over a distribution of decision trees induced by bootstrap sampling and feature randomness. Instead of finite summation, RF approximates E[T(x)] where T is a random tree function. As n_estimators → ∞, the estimator converges to this expectation, reducing variance while preserving structural bias.