def subsample(self, by, value):
""" Returns a subsampled RDMs with repetitions if values are repeated
Args:
by(String): the descriptor by which the subset selection
is made from descriptors
value: the value by which the subset selection is made
from descriptors
Returns:
RDMs object, with subsampled RDMs
"""
if by is None:
by = 'index'
if (
type(value) is list or
type(value) is tuple or
type(value) is np.ndarray):
selection = [np.asarray(self.rdm_descriptors[by] == i).nonzero()[0]
for i in value]
selection = np.concatenate(selection)
else:
selection = np.where(self.rdm_descriptors[by] == value)
dissimilarities = self.dissimilarities[selection, :]
descriptors = self.descriptors
pattern_descriptors = self.pattern_descriptors
rdm_descriptors = extract_dict(self.rdm_descriptors, selection)
dissimilarity_measure = self.dissimilarity_measure
rdms = RDMs(dissimilarities=dissimilarities,
descriptors=descriptors,
rdm_descriptors=rdm_descriptors,
pattern_descriptors=pattern_descriptors,
dissimilarity_measure=dissimilarity_measure)
return rdms
def subset(self, by, value):
""" Returns a set of fewer RDMs matching descriptor values
Args:
by(String): the descriptor by which the subset selection
is made from descriptors
value: the value by which the subset selection is made
from descriptors
Returns:
RDMs object, with fewer RDMs
"""
if by is None:
by = 'index'
selection = bool_index(self.rdm_descriptors[by], value)
dissimilarities = self.dissimilarities[selection, :]
descriptors = self.descriptors
pattern_descriptors = self.pattern_descriptors
rdm_descriptors = extract_dict(self.rdm_descriptors, selection)
dissimilarity_measure = self.dissimilarity_measure
rdms = RDMs(dissimilarities=dissimilarities,
descriptors=descriptors,
rdm_descriptors=rdm_descriptors,
pattern_descriptors=pattern_descriptors,
dissimilarity_measure=dissimilarity_measure)
return rdms
def subset_pattern(self, by, value):
""" Returns a smaller RDMs with patterns with certain descriptor values
Args:
by(String): the descriptor by which the subset selection
is made from pattern_descriptors
value: the value by which the subset selection is made
from pattern_descriptors
Returns:
RDMs object, with fewer patterns
"""
if by is None:
by = 'index'
selection = bool_index(self.pattern_descriptors[by], value)
dissimilarities = self.get_matrices()[:, selection][:, :, selection]
descriptors = self.descriptors
pattern_descriptors = extract_dict(
self.pattern_descriptors, selection)
rdm_descriptors = self.rdm_descriptors
dissimilarity_measure = self.dissimilarity_measure
rdms = RDMs(dissimilarities=dissimilarities,
descriptors=descriptors,
rdm_descriptors=rdm_descriptors,
pattern_descriptors=pattern_descriptors,
dissimilarity_measure=dissimilarity_measure)
return rdms
def subsample_pattern(self, by, value):
""" Returns a subsampled RDMs with repetitions if values are repeated
This function now generates Nans where the off-diagonal 0s would
appear. These values are trivial to predict for models and thus
need to be marked and excluded from the evaluation.
Args:
by(String): the descriptor by which the subset selection
is made from descriptors
value: the value by which the subset selection is made
from descriptors
Returns:
RDMs object, with subsampled patterns
"""
if by is None:
by = 'index'
if (
type(value) is list or
type(value) is tuple or
type(value) is np.ndarray):
desc = self.pattern_descriptors[by]
selection = [np.asarray(desc == i).nonzero()[0]
for i in value]
selection = np.concatenate(selection)
else:
selection = np.where(self.rdm_descriptors[by] == value)
selection = np.sort(selection)
dissimilarities = self.get_matrices()
for i_rdm in range(self.n_rdm):
np.fill_diagonal(dissimilarities[i_rdm], np.nan)
selection = np.sort(selection)
dissimilarities = dissimilarities[:, selection][:, :, selection]
descriptors = self.descriptors
pattern_descriptors = extract_dict(
self.pattern_descriptors, selection)
rdm_descriptors = self.rdm_descriptors
dissimilarity_measure = self.dissimilarity_measure
rdms = RDMs(dissimilarities=dissimilarities,
descriptors=descriptors,
rdm_descriptors=rdm_descriptors,
pattern_descriptors=pattern_descriptors,
dissimilarity_measure=dissimilarity_measure)
return rdms
def subset_channel(self, by, value):
""" Returns a subsetted Dataset defined by certain channel value
Args:
by(String): the descriptor by which the subset selection is
made from channel dimension
value: the value by which the subset selection is made
from channel dimension
Returns:
Dataset, with subset defined by the selected channel_descriptor
"""
selection = bool_index(self.channel_descriptors[by], value)
measurements = self.measurements[:, selection]
descriptors = self.descriptors
obs_descriptors = self.obs_descriptors
channel_descriptors = extract_dict(self.channel_descriptors, selection)
dataset = Dataset(measurements=measurements,
descriptors=descriptors,
obs_descriptors=obs_descriptors,
channel_descriptors=channel_descriptors)
return dataset
def split_obs(self, by):
""" Returns a list Datasets splited by obs
Args:
by(String): the descriptor by which the splitting is made
Returns:
list of Datasets, splitted by the selected obs_descriptor
"""
unique_values = get_unique_unsorted(self.obs_descriptors[by])
dataset_list = []
for v in unique_values:
selection = (self.obs_descriptors[by] == v)
measurements = self.measurements[selection, :]
descriptors = self.descriptors
obs_descriptors = extract_dict(self.obs_descriptors, selection)
channel_descriptors = self.channel_descriptors
dataset = Dataset(measurements=measurements,
descriptors=descriptors,
obs_descriptors=obs_descriptors,
channel_descriptors=channel_descriptors)
dataset_list.append(dataset)
return dataset_list