Make predictions based on prototypes

Share:

Description

Performs nearest-prototype classification like predict.protoclass, except only requires the ntest-by-m matrix of distances between test features and z rather than requiring these matrices themselves. (Using this saves repeated computation of the pairwise distances...)

Usage

1
  predictwithd.protoclass(object, d.test.z)

Arguments

object

output of protoclass function

d.test.z

ntest-by-m matrix of distances between test features and z

See Also

predict.protoclass