imputegap.algorithms.zero_impute package¶

The imputegap.algorithms.zero_impute package contains various imputation algorithms used for handling missing values in time series data. This package supports multiple imputation techniques like CDRec, MRNN, IIM, and more.

Submodules¶

Modules¶

imputegap.algorithms.zero_impute.zero_impute(incomp_data, params=None)[source]¶

Impute missing values (NaNs) with zeros in the time series.

Parameters¶

incomp_datanumpy.ndarray: The input time series matrix with missing values represented as NaNs.
paramsdict, optional: Optional parameters for the algorithm. This is not used in the current implementation but can be passed for future extensions (default is None).

Returns¶

numpy.ndarray: The imputed matrix where all NaN values have been replaced by zeros.

Notes¶

This simple imputation strategy replaces all missing values (NaNs) with zeros. This can be useful for initializing datasets where more complex imputation methods will follow.

Example¶

>>> incomp_data = np.array([[1, 2, np.nan], [4, np.nan, 6]])
>>> recov_data = zero_impute(incomp_data)
>>> print(recov_data)
array([[1., 2., 0.],
       [4., 0., 6.]])

author:: Quentin Nater