object of class classiKernel to get predictions from
(optional) new data to predict from with observations as rows. Do not derive this data,
this will be done automatically if required by the model. If NULL,
the training data is predicted, currently without using a leave-one-out prediction.
one of 'response' or 'prob', indicating the type of prediction. Choose
'response' to return a vector of length nrow(newdata) containing the
most predicted class.
Choose 'prob' to return a matrix with nrow(newdata) rows containing
the probabilities for the classes as columns.
Should the prediction be parallelized?
Uses parallelMap for
parallelization. See ... for further arguments.