def test_from_columns(self):
tsn = "TEST_TIME_SERIES"
fset = FeatureExtractionSettings()
self.assertRaises(TypeError, fset.from_columns, 42)
self.assertRaises(TypeError, fset.from_columns, 42)
self.assertRaises(ValueError, fset.from_columns, ["This is not a column name"])
self.assertRaises(ValueError, fset.from_columns, ["This__neither"])
self.assertRaises(ValueError, fset.from_columns, ["This__also__not"])
# Aggregate functions
feature_names = [tsn + '__sum_values', tsn + "__median", tsn + "__length", tsn + "__sample_entropy"]
# Aggregate functions with params
feature_names += [tsn + '__quantile__q_10', tsn + '__quantile__q_70', tsn + '__number_peaks__n_30',
tsn + '__value_count__value_inf', tsn + '__value_count__value_-inf',
tsn + '__value_count__value_nan']
# Apply functions
feature_names += [tsn + '__ar_coefficient__k_20__coeff_4', tsn + '__ar_coefficient__coeff_10__k_-1']
cset = fset.from_columns(feature_names)
six.assertCountEqual(self, list(cset.kind_to_calculation_settings_mapping[tsn].keys()),
["sum_values", "median", "length", "sample_entropy", "quantile", "number_peaks", "ar_coefficient",
"value_count"])
self.assertEqual(cset.kind_to_calculation_settings_mapping[tsn]["sum_values"], None)
self.assertEqual(cset.kind_to_calculation_settings_mapping[tsn]["ar_coefficient"],
[{"k": 20, "coeff": 4}, {"k": -1, "coeff": 10}])
self.assertEqual(cset.kind_to_calculation_settings_mapping[tsn]["value_count"],
[{"value": np.PINF}, {"value": np.NINF}, {"value": np.NaN}])
def transform(self, X):
"""
After the fit step, it is known which features are relevant. Only extract those from the time series handed in
with the function :func:`~set_timeseries_container`.
If evaluate_only_added_features is False, also delete the irrelevant, already present features in the data frame.
:param X: the data sample to add the relevant (and delete the irrelevant) features to.
:type X: pandas.DataFrame or numpy.array
:return: a data sample with the same information as X, but with added relevant time series features and
deleted irrelevant information (only if evaluate_only_added_features is False).
:rtype: pandas.DataFrame
"""
if self.feature_selector.relevant_features is None:
raise RuntimeError("You have to call fit before.")
if self.timeseries_container is None:
raise RuntimeError(
"You have to provide a time series using the set_timeseries_container function before."
)
self.feature_extractor.set_timeseries_container(
self.timeseries_container)
relevant_time_series_features = set(
self.feature_selector.relevant_features) - set(
pd.DataFrame(X).columns)
relevant_extraction_settings = FeatureExtractionSettings.from_columns(
relevant_time_series_features)
relevant_extraction_settings.set_default = False
# Set imputing strategy
if self.feature_extractor.settings.IMPUTE is impute_dataframe_range:
relevant_extraction_settings.IMPUTE = partial(
impute_dataframe_range,
col_to_max=self.col_to_max,
col_to_min=self.col_to_min,
col_to_median=self.col_to_median)
else:
relevant_extraction_settings.IMPUTE = self.feature_extractor.settings.IMPUTE
relevant_feature_extractor = FeatureAugmenter(
settings=relevant_extraction_settings,
column_id=self.feature_extractor.column_id,
column_sort=self.feature_extractor.column_sort,
column_kind=self.feature_extractor.column_kind,
column_value=self.feature_extractor.column_value)
relevant_feature_extractor.set_timeseries_container(
self.feature_extractor.timeseries_container)
X_augmented = relevant_feature_extractor.transform(X)
return X_augmented.copy().loc[:,
self.feature_selector.relevant_features]
def transform(self, X):
"""
After the fit step, it is known which features are relevant. Only extract those from the time series handed in
with the function :func:`~set_timeseries_container`.
If evaluate_only_added_features is False, also delete the irrelevant, already present features in the data frame.
:param X: the data sample to add the relevant (and delete the irrelevant) features to.
:type X: pandas.DataFrame or numpy.array
:return: a data sample with the same information as X, but with added relevant time series features and
deleted irrelevant information (only if evaluate_only_added_features is False).
:rtype: pandas.DataFrame
"""
if self.feature_selector.relevant_features is None:
raise RuntimeError("You have to call fit before.")
if self.timeseries_container is None:
raise RuntimeError(
"You have to provide a time series using the set_timeseries_container function before."
)
self.feature_extractor.set_timeseries_container(
self.timeseries_container)
# We can only extract features that originate from time series
relevant_extraction_settings = FeatureExtractionSettings.from_columns(
list(
set(self.feature_selector.relevant_features) -
set(pd.DataFrame(X).columns)))
relevant_extraction_settings.set_default = False
relevant_extraction_settings.IMPUTE = self.feature_extractor.settings.IMPUTE
feature_augmenter_restricted = FeatureAugmenter(
settings=relevant_extraction_settings,
column_id=self.feature_extractor.column_id,
column_sort=self.feature_extractor.column_sort,
column_kind=self.feature_extractor.column_kind,
column_value=self.feature_extractor.column_value)
feature_augmenter_restricted.set_timeseries_container(
self.feature_extractor.timeseries_container)
if self.evaluate_only_added_features:
X_tsfresh = feature_augmenter_restricted.transform(
X).loc[:, self.feature_selector.relevant_features]
return pd.concat([X_tsfresh, X], axis=1)
else:
X_tsfresh = feature_augmenter_restricted.transform(X)
return X_tsfresh.loc[:, self.feature_selector.relevant_features]