nemos.basis._basis.Basis.setup_basis

abstractmethod Basis.setup_basis(*xi)

Pre-compute all basis state variables.

This method is intended to be equivalent to the sklearn transformer fit method. As the latter, it computes all the state attributes, and store it with the convention that the attribute name must end with “_”, for example self.kernel_, self._input_shape_.

The method differs from transformer’s fit for the structure of the input that it accepts. In particular, _fit_basis accepts a number of different time series, one per 1D basis component, while fit requires all inputs to be concatenated in a single array.

Return type:

TsdFrame | ndarray[tuple[int, ...], dtype[TypeVar(_ScalarType_co, bound= generic, covariant=True)]]

Parameters:

xi (ArrayLike)