sisireg (version 1.1.2)

ssrmlp_train: 2-layer MLP with partial sum optimization

Description

Calculates the weights of a 2-layer MLP with respect to the partial sums critereon

Usage

ssrmlp_train(X, Y, std=TRUE, opt='ps', hl = NULL, W = NULL,
  k=10, fn=4, eta=0.75, maxIter=1000,
  facfct_ex = NULL, errfct_ex = NULL, alpha = NULL)

Value

W

List with weight matrices.

Arguments

X

matrix with n-dimensional coordinates.

Y

array with observations.

std

optional: standardizing values if TRUE.

opt

optional: optimizing function ('ps', 'lse', 'ps_l1', 'ps_lse', 'ext').

hl

optional: array tupel with number of perceptrons in each layer.

W

optional: previously calculates weights for refining the model.

k

optional: number of neighbors per quadrant.

fn

optional: quantile for partial sums.

eta

optional: constant factor of the gradient algorithm.

maxIter

optional: number of iterations for the numeric solver.

facfct_ex

optional: first derivative of external error function, for opt='ext' only.

errfct_ex

optional: external error function, for opt='ext' only.

alpha

optional: weight parameter for error function.

Author

Dr. Lars Metzner

References

Dr. Lars Metzner (2021) Adäquates Maschinelles Lernen. Independently Published.

Examples

# \donttest{
# generate data
set.seed(42)
x <- rnorm(300)
y <- rnorm(300)
z <- rnorm(300) + atan2(x, y)
# coordinates
X <- matrix(cbind(x,y), ncol = 2)
Y <- as.double(z)
# Training
W <- ssrmlp_train(X, Y)
# }