Est. Budget: $200.00
Extend the gbm ("gradient boosting machine"/"generalised boosting machine") to output an additional component in the gbm.object returned by gbm.
This would be a logical matrix with a number of rows equal to the number of training observations in the data, and a number of columns equal to the number of boosting iterations.
A matrix cell would be True if the training observation corresponding to that row was used in training the ensemble element corresponding to that column, and ...