get_period_shift: Period shift
| get_period_shift | R Documentation |
Period shift
Description
Period shift
Usage
get_period_shift(timesteps = 1L, lag = 0L, type = "univariate")
Arguments
timesteps |
The number of timesteps. |
lag |
The number of considered lags on feature side. |
type |
The type of time series: |
Details
The period shift denotes the number of past periods starting from a certain period t, whose values of X are needed to describe Y in period t and which cannot be used to extract Y values. This number of past periods depends on the type of timeseries (univariate or multivariate), the number of timesteps and the underpinned number of lags. In other words, the period shift is the number of periods to go backwards to get features (X) for outcomes (Y) or as a question: how many samples get lost respectively must be used as X for Y description/prediction?
Value
The period shift.
See Also
Other Recurrent Neural Network (RNN):
as_LSTM_X(),
as_LSTM_Y(),
as_LSTM_data_frame(),
as_LSTM_period_outcome(),
as_lag(),
as_timesteps(),
build_LSTM(),
fit_LSTM(),
get_LSTM_XY(),
load_weights_ANN(),
predict_ANN(),
save_weights_ANN(),
start_invert_differencing()
