# Are Random Forests Turing-Complete?

Is the random forests algorithm Turing-complete? As in, can any algorithm be represented by a given "tree" in the forest?

Random forests are useful only for inputs of a given size, so they are not really eligible for Turing-completeness. However, any function on inputs of size $n$ can be computed by a decision tree, so in that sense the model is complete (though in practice we limit the depth of the trees).