kedm.eval_smap

kedm.eval_smap(lib: Annotated[numpy.typing.ArrayLike, numpy.float32], pred: Annotated[numpy.typing.ArrayLike, numpy.float32], *, target: Annotated[numpy.typing.ArrayLike, numpy.float32] = None, E: SupportsInt = 1, tau: SupportsInt = 1, Tp: SupportsInt = 1, theta: SupportsFloat = 1.0) → float

Predict a time series from another using S-Map and quantify its predictive skill.

Parameters:

  • lib – Library time series

  • pred – Prediction time series

  • target – Target time series (defaults to pred if None)

  • E – Embedding dimension

  • tau – Time delay

  • Tp – Prediction interval

  • theta – Neighbor localization exponent

Returns:

Pearson’s correlation coefficient between observation and prediction

Note

If target is given (cross mapping), the prediction is compared to target. Otherwise, the prediction is compared to pred.