Fits various tracks (different random starting values) and chooses best model using torch

Fits n_init tracks with different initial values and decides on best model based on information criteria.

Usage

nn_fit_torch(
  x,
  y,
  q,
  n_init,
  inf_crit = "BIC",
  response = "continuous",
  unif = 3,
  maxit = 1000,
  lambda = 0,
  min_delta = 1e-08,
  patience = 10,
  batch_size = 32,
  ...
)

Arguments

x

Matrix of covariates

y

Vector of response

q

Number of hidden nodes

n_init

Number of random initialisations (tracks)

inf_crit

Information criterion: "BIC" (default), "AIC" or "AICc"

response

"continuous" (default) or "binary"

unif

Random initial values max value

maxit

maximum number of iterations for nnet (default = 100)

lambda

ridge penalty

min_delta

tolerance for early stopping

patience

patience for ealy stopping

batch_size

batch size

...

additional argument for nnet

Value

The best model from the different tracks