A Random Forest works by aggregating the results of many decision trees. Recently, I was reading about how the RandomForest aggregates the results, and it made me question whether the results from Decision trees are a single class prediction or a probability of each class.
the documentation for
"The predicted class probabilities of an input sample is computed as the mean predicted class probabilities of the trees in the forest. The class probability of a single tree is the fraction of samples of the same class in a leaf."
the part about "mean predicted class probabilities" indicates that the decision trees are non-deterministic. Furthermore, the lecture by Nando De Freitas here also talks of class probabilities at around 30 minutes.
My question is - how is it possible to get class probabilities from a single decision tree?
As far as I know, the default for a RandomForestClassifier in sklearn is a deterministic decision tree (for something like ExtraTreesClassifier, that is not the case). So, where do these said "class probabilities" for a single tree come from?