knn-sampler is a non parametrical conditional density function estimator and sampler. It is very useful when you need to estimate errors in a regression problem, since instead of returning the regressor pointiwse prediction, it returns the predicted values distribution.
two estimators are avalible:
[x] KNNSampler - Based o KNN search algorithm (unsupervised) [x] ForestSampler - Based on Forest algorithms (supervised)