predict.KDSNensembleDisk: Predict kernel deep stacking networks ensembles...
In kernDeepStackNet: Kernel Deep Stacking Networks

Description Usage Arguments Details Value Note Author(s) See Also

Predicts new data with a given kernel deep stacking network ensembles. The model is not stored in workspace, but on disk in temporary folder, where it has been created. The temporary file folder should be available in the working directory. Note that this function is still experimental.

1 2	## S3 method for class 'KDSNensembleDisk' predict(object, newx, ...)

`object`	Object of class `KDSNensemble`. This object is generated with the function `fitEnsembleKDSN`.
`newx`	New data design matrix, for which predictions are needed. Variables must be in the same order, as the original training data.
`...`	Further arguments to `predict` function.

The data is put through all specified layers of the kernel deep stacking network. The weights are not random, but fixed at the values generated by the fitting process. Examples are given in the help page of fitEnsembleKDSN.

A prediction matrix will be returned. Each row corresponds to one observation and each column is another KDSN ensemble.

Do not rename the temporary directory. Otherwise the model files will not be found.

Thomas Welchowski welchow@imbie.meb.uni-bonn.de

fitKDSN, fitEnsembleKDSN

kernDeepStackNet documentation built on May 2, 2019, 8:16 a.m.