Neural Network Binary Time Series Forecasts

Feed-forward neural networks with a single hidden layer and lagged inputs for forecasting univariate binary time series.

Usage

BINNET(formula, n_nodes = NULL, n_networks = 20, scale_inputs = TRUE, ...)

Arguments

formula

Model specification (see "Specials" section).

n_nodes

Number of nodes in the hidden layer. Default is half of the number of external regressors plus 1.

n_networks

Number of networks to fit with different random starting weights. These are then averaged when producing forecasts.

scale_inputs

If TRUE, inputs are scaled by subtracting the column means and dividing by their respective standard deviations.

...

Other arguments passed to \link[nnet]{nnet}.

Value

A model specification.

Details

A feed-forward neural network is fitted with a single hidden layer containing size nodes.

Exogenous regressors are used as inputs. A total of repeats networks are fitted, each with random starting weights. These are then averaged when computing forecasts.

Examples

melb_rain |>
  model(nn = BINNET(Wet ~ fourier(K = 1, period = "year")))
#> # A mable: 1 x 1
#>            nn
#>       <model>
#> 1 <BINNET: 2>