as_LSTM_period_outcome: Rebuild data frame

In stschn/deepANN: Neural Network Toolbox

Rebuild data frame

Description

Extract columns from a data frame under consideration of time series type, timesteps and lags.

Usage

as_LSTM_period_outcome(
  dataset,
  columns,
  timesteps = 1L,
  lag = 0L,
  type = "univariate"
)

Arguments

dataset	A data set, usually a matrix or data frame with training or test data.
columns	The names or indices of the columns which should be extracted.
timesteps	The number of timesteps; stands for the number of different periods within one sample (record) of the resampled feature matrix, returned by as_LSTM_X.
lag	The number of considered lags on feature side.
type	The type of time series: univariate or multivariate.

Details

Within time series analysis, get_LSTM_XY extracts X and Y of a matrix or data frame in a LSTM compatible preformat. Then, these values are resampled thru as_LSTM_X and as_LSTM_Y. During extraction, the period shift is calculated to determine the number of past periods which are used to extract X values for Y description/prediction. Because of this number of periods cannot be used to extract Y values, this number must be deleted from dataset.

Value

A data frame with columns that can be used for quality assurance or for graphical representation together with the predicted values produced by predict_LSTM.

See Also

get_LSTM_XY, get_period_shift, predict_ANN.

Other Recurrent Neural Network (RNN): as_LSTM_X(), as_LSTM_Y(), as_LSTM_data_frame(), as_lag(), as_timesteps(), build_LSTM(), fit_LSTM(), get_LSTM_XY(), get_period_shift(), load_weights_ANN(), predict_ANN(), save_weights_ANN(), start_invert_differencing()

stschn/deepANN documentation built on June 25, 2024, 7:27 a.m.