Secure your code as it's written. Use Snyk Code to scan source code in minutes - no build needed - and fix issues immediately.
tlc_split_fn = time_learning_curve_splitter(training_time_limit='2016-01-01', time_column='time', min_samples=0)
sc_split_fn = stability_curve_time_splitter(training_time_limit='2016-01-01', time_column='time', min_samples=0)
fw_sc_split_fn = forward_stability_curve_time_splitter(
training_time_start="2015-01-01",
training_time_end="2016-01-01",
holdout_gap=timedelta(days=30),
holdout_size=timedelta(days=30),
step=timedelta(days=30),
time_column='time'
)
# Validate results
cv_results = validator(df, cv_split_fn, train_fn, eval_fn)['validator_log']
tlc_results = validator(df, tlc_split_fn, train_fn, eval_fn)['validator_log']
sc_results = validator(df, sc_split_fn, train_fn, eval_fn)['validator_log']
fw_sc_results = validator(df, fw_sc_split_fn, train_fn, eval_fn)['validator_log']
# temporal evaluation results
predict_fn, _, _ = train_fn(df)
temporal_week_results = temporal_week_eval_fn(predict_fn(df))
temporal_year_results = temporal_year_eval_fn(predict_fn(df))
# Define extractors
base_extractors = combined_evaluator_extractor(base_extractors=[
evaluator_extractor(evaluator_name="r2_evaluator__target"),
evaluator_extractor(evaluator_name="spearman_evaluator__target")
])
splitter_extractor = split_evaluator_extractor(split_col='RAD', split_values=[4.0, 5.0, 24.0],
def test_validator():
result = validator(data, split_fn, train_fn, eval_fn, perturb_fn_train, perturb_fn_test)
validator_log = result["validator_log"]
assert len(validator_log) == 2
assert validator_log[0]['fold_num'] == 0
assert result['train_log']['xgb_classification_learner']['features'] == ['f1']
assert len(validator_log[0]['eval_results']) == 3
assert validator_log[1]['fold_num'] == 1
assert len(validator_log[1]['eval_results']) == 1
perturbator_log = result["perturbator_log"]
assert perturbator_log['perturbated_train'] == []
assert perturbator_log['perturbated_test'] == ['rows']
sc_split_fn = stability_curve_time_splitter(training_time_limit='2016-01-01', time_column='time', min_samples=0)
fw_sc_split_fn = forward_stability_curve_time_splitter(
training_time_start="2015-01-01",
training_time_end="2016-01-01",
holdout_gap=timedelta(days=30),
holdout_size=timedelta(days=30),
step=timedelta(days=30),
time_column='time'
)
# Validate results
cv_results = validator(df, cv_split_fn, train_fn, eval_fn)['validator_log']
tlc_results = validator(df, tlc_split_fn, train_fn, eval_fn)['validator_log']
sc_results = validator(df, sc_split_fn, train_fn, eval_fn)['validator_log']
fw_sc_results = validator(df, fw_sc_split_fn, train_fn, eval_fn)['validator_log']
# temporal evaluation results
predict_fn, _, _ = train_fn(df)
temporal_week_results = temporal_week_eval_fn(predict_fn(df))
temporal_year_results = temporal_year_eval_fn(predict_fn(df))
# Define extractors
base_extractors = combined_evaluator_extractor(base_extractors=[
evaluator_extractor(evaluator_name="r2_evaluator__target"),
evaluator_extractor(evaluator_name="spearman_evaluator__target")
])
splitter_extractor = split_evaluator_extractor(split_col='RAD', split_values=[4.0, 5.0, 24.0],
base_extractor=base_extractors)
tlc_split_fn = time_learning_curve_splitter(training_time_limit='2016-01-01', time_column='time', min_samples=0)
sc_split_fn = stability_curve_time_splitter(training_time_limit='2016-01-01', time_column='time', min_samples=0)
fw_sc_split_fn = forward_stability_curve_time_splitter(
training_time_start="2015-01-01",
training_time_end="2016-01-01",
holdout_gap=timedelta(days=30),
holdout_size=timedelta(days=30),
step=timedelta(days=30),
time_column='time'
)
# Validate results
cv_results = validator(df, cv_split_fn, train_fn, eval_fn)['validator_log']
tlc_results = validator(df, tlc_split_fn, train_fn, eval_fn)['validator_log']
sc_results = validator(df, sc_split_fn, train_fn, eval_fn)['validator_log']
fw_sc_results = validator(df, fw_sc_split_fn, train_fn, eval_fn)['validator_log']
# temporal evaluation results
predict_fn, _, _ = train_fn(df)
temporal_week_results = temporal_week_eval_fn(predict_fn(df))
temporal_year_results = temporal_year_eval_fn(predict_fn(df))
# Define extractors
base_extractors = combined_evaluator_extractor(base_extractors=[
evaluator_extractor(evaluator_name="r2_evaluator__target"),
evaluator_extractor(evaluator_name="spearman_evaluator__target")
])
splitter_extractor = split_evaluator_extractor(split_col='RAD', split_values=[4.0, 5.0, 24.0],
base_extractor=base_extractors)
sc_split_fn = stability_curve_time_splitter(training_time_limit='2016-01-01', time_column='time', min_samples=0)
fw_sc_split_fn = forward_stability_curve_time_splitter(
training_time_start="2015-01-01",
training_time_end="2016-01-01",
holdout_gap=timedelta(days=30),
holdout_size=timedelta(days=30),
step=timedelta(days=30),
time_column='time'
)
# Validate results
cv_results = validator(df, cv_split_fn, train_fn, eval_fn)['validator_log']
tlc_results = validator(df, tlc_split_fn, train_fn, eval_fn)['validator_log']
sc_results = validator(df, sc_split_fn, train_fn, eval_fn)['validator_log']
fw_sc_results = validator(df, fw_sc_split_fn, train_fn, eval_fn)['validator_log']
# temporal evaluation results
predict_fn, _, _ = train_fn(df)
temporal_week_results = temporal_week_eval_fn(predict_fn(df))
temporal_year_results = temporal_year_eval_fn(predict_fn(df))
# Define extractors
base_extractors = combined_evaluator_extractor(base_extractors=[
evaluator_extractor(evaluator_name="r2_evaluator__target"),
evaluator_extractor(evaluator_name="spearman_evaluator__target")
])
splitter_extractor = split_evaluator_extractor(split_col='RAD', split_values=[4.0, 5.0, 24.0],
base_extractor=base_extractors)
temporal_week_splitter_extractor = temporal_split_evaluator_extractor(