I'm working with the mldr::birds dataset and I have created two birds_training and a birds_testing datasets for calibration and for prediction performance, respectively.
After fitting the rFerns model on the birds_training with dimensions of 451(rows) x 279 (columns), I make a prediction on the new dataset passed as an argument "newdata" and with dimensions of 194(rows) x 279 (columns).
Everything works great, except the dimensions of the prediction has the same number of instances (row) of the birds_training (451) and no the number of instances (rows) from the birds_testing (194) Data sets.
Many thanks for your Help!
Seems like you haven't passed the testing set to the predict function, this way it just drops the OOB predictions -- such use case in covered by unit tests, so bug is unlikely.
Anyway, please provide minimal reproducible example.