smooth.ES.states

property ES.states: ndarray[tuple[int, ...], dtype[_ScalarType_co]]

$states).

The state matrix stores the evolution of all model components including level, trend (if present), seasonal components (if present), and ARIMA states (if present). Each column represents a different state component, and each row represents a time point.

Returns:

2D array of shape (n_states, obs_in_sample + 1). Columns represent level, trend (if model has trend), and seasonal components (if model has seasonality, one column per lag).

Return type:

NDArray

Raises:

ValueError – If the model has not been fitted yet.

Examples

>>> model = ADAM(model="AAA", lags=12)
>>> model.fit(y)
>>> states = model.states
>>> level = states[0, :]  # Level component over time

Type:

State matrix containing component values over time (R