denspp.offline.data_augmentation.downsampler#

Module Contents#

Functions#

augmentation_downsampling

Function for data augmentation using downsampling.

API#

denspp.offline.data_augmentation.downsampler.augmentation_downsampling(data: numpy.ndarray, label: numpy.ndarray, n_downsampling: int, drop_samples: bool = False) → tuple[numpy.ndarray, numpy.ndarray][source]#

Function for data augmentation using downsampling.

Parameters:

  • data – Numpy array with data from dataset with shape [num_samples, num_ch (optional), num_timepoints]

  • label – Numpy array with labels from dataset with shape [num_samples, ]

  • n_downsampling – Integer with level for downsampling (0: error, 1: no change, 2: reduced by 2, 3: reduced by 3, …)

  • drop_samples – Boolean, whether to drop samples (true) or augment samples (false)

Returns:

Tuple with data and label for the new augmented dataset using downsampling.