fourierTransPredict: Prediction based on random Fourier transformation

Description Usage Arguments Value Note Author(s) References See Also Examples

Description

Based on prior given weights, the estimated Fourier transformation is applied to new data.

Usage

1

Arguments

newx

New data design matrix.

rW

Prior drawn random weight matrix.

Value

Numeric transformed data matrix with dimension 2*Dim x n.

Note

This function is not intended to be called directly by the user. Should only be used by experienced users, who want to customize the model. It is called in the estimation process of the kernel deep stacking network, e.g. fitKDSN.

Author(s)

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

References

Po-Seng Huang and Li Deng and Mark Hasegawa-Johnson and Xiaodong He, (2013), Random Features for kernel deep convex network, Proceedings IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

See Also

randomFourierTrans

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
# Generate data matrix
set.seed(50)
X <- matrix(rnorm(100*3), ncol=3)

# Apply a random Fourier transformation of higher dimension
rft <- randomFourierTrans(X=X, Dim=3, sigma=1, seedW=0)

# New data matrix
set.seed(100)
Xnew <- matrix(rnorm(100*3), ncol=3)

# Apply same Fourier transformation on new data
newZ <- fourierTransPredict(newx=Xnew, rW=rft$rW)
head(newZ)

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