def _fit_single(self, model, X, y, masked_data=None):
Validation.check_dataset(X, y)
if len(X) != 0:
# Pre-compute target outputs if none are passed.
if masked_data is None:
output_dim = Validation.get_output_dimension(y)
masked_data = self.masking_operation.get_predictions_after_masking(self.explained_model, X, y,
batch_size=
self.model_builder.batch_size,
downsample_factors=
self.downsample_factors,
flatten=
self.flatten_for_explained_model)
masked_data = TensorflowCXPlain._clean_output_dims(output_dim, masked_data)
self.last_masked_data = masked_data
if self.model_filepath is None:
from tempfile import NamedTemporaryFile
model_filepath = NamedTemporaryFile(delete=False).name
else:
model_filepath = self.model_filepath
self.last_history = self.model_builder.fit(model, masked_data, y, model_filepath)
return self
def _build_model(self, X, y):
Validation.check_dataset(X, y)
if Validation.is_variable_length(X):
raise ValueError("Variable length inputs to CXPlain are currently not supported.")
n, p = Validation.get_input_dimension(X)
output_dim = Validation.get_output_dimension(y)
if self.model is None:
if self.num_models == 1:
build_fun = self._build_single
else:
build_fun = self._build_ensemble
self.model, self.prediction_model = build_fun(input_dim=p, output_dim=output_dim)
def score(self, X, y, sample_weight=None, masked_data=None):
"""
Evaluates the performance, in terms of causal loss, of the current CXPlain model
:param X: The data samples to be evaluated. The first dimension must be the number of samples. (Required)
:param y: The ground truth labels to be compared to. The first dimension must be the number of
samples. (Required)
:param sample_weight: The sample weight to apply to the samples in X during evaluation. The first dimension
must be the number of samples and it must match that of __X__ and __y__.
If None, equal weihting is used (Optional, default: None).
:param masked_data: An array of precomputed masked data as can be obtained from __get_masked_data__.
If None, the masked data is computed. If set, the precomputed masked data is used for
scoring and computation of the masked data is skipped (Optional, default: None).
:return: Score results as returned by self.model_builder.evaluate(model, X, y, sample_weight) either
(i) as a single score result if __num_models__ = 1 or as a list of score results
if __num_models__ is greater than 1.
:exception AssertionError Thrown if the explanation model has not been fitted using __fit__ yet.
"""
if self.model is None:
raise AssertionError("Model must be initialised when calling __predict__. "
"Did you forget to __fit__ the explanation model?")
output_dim = Validation.get_output_dimension(y)
if masked_data is None:
masked_data = self.masking_operation.get_predictions_after_masking(self.explained_model, X, y,
batch_size=
self.model_builder.batch_size,
downsample_factors=
self.downsample_factors,
flatten=
self.flatten_for_explained_model)
masked_data = TensorflowCXPlain._clean_output_dims(output_dim, masked_data)
self.last_masked_data = masked_data
if self.num_models == 1:
return_value = self._score_single(self.model, masked_data, y, sample_weight)
else:
return_value = [self._score_single(model, masked_data, y, sample_weight) for model in self.model]
return return_value