Skip to content

Fix bug in separate by gap for time series #3256

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
freddyaboulton merged 4 commits into from
Jan 20, 2022
Merged

Fix bug in separate by gap for time series #3256

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
219dc3e
Fix bug in separate by gap
freddyaboulton Jan 19, 2022
9d88ea5
Add to release notes
freddyaboulton Jan 19, 2022
eae8013
Undo notebook changes
freddyaboulton Jan 19, 2022
fbbcf00
Fix bug in release notes
freddyaboulton Jan 19, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions docs/source/release_notes.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
* Removed potential prediction explanations failure when pipelines predicted a class with probability 1 :pr:`3221`
* Dropped NaNs before partial dependence grid generation :pr:`3235`
* Fixed bug where ``InvalidTargetDataCheck`` would not check time series regression targets :pr:`3251`
* Fixed bug in ``are_datasets_separated_by_gap_time_index`` :pr:`3256`
* Changes
* Raised lowest compatible numpy version to 1.21.0 to address security concerns :pr:`3207`
* Changed the default objective to ``MedianAE`` from ``R2`` for time series regression :pr:`3205`
Expand Down
15 changes: 15 additions & 0 deletions evalml/tests/utils_tests/test_gen_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -837,3 +837,18 @@ def test_time_series_pipeline_validates_holdout_data(
assert (
result.error_codes[0] == ValidationErrorCode.INVALID_HOLDOUT_GAP_SEPARATION
)


def test_year_start_separated_by_gap():
X = pd.DataFrame(
{
"time_index": pd.Series(
pd.date_range("1960-01-01", freq="AS-JAN", periods=35)
)
}
)
train = X.iloc[:30]
test = X.iloc[32:36]
assert are_datasets_separated_by_gap_time_index(
train, test, {"time_index": "time_index", "gap": 2}
)
9 changes: 2 additions & 7 deletions evalml/utils/gen_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@

import numpy as np
import pandas as pd
from pandas.tseries.frequencies import to_offset
from sklearn.utils import check_random_state

from evalml.exceptions import MissingComponentError, ValidationErrorCode
Expand Down Expand Up @@ -611,13 +612,7 @@ def are_datasets_separated_by_gap_time_index(train, test, pipeline_params):

first_testing_date = test_copy[test_copy.ww.time_index].iloc[0]
last_training_date = train_copy[train_copy.ww.time_index].iloc[-1]
dt_difference = first_testing_date - last_training_date

try:
units_difference = dt_difference / freq
except ValueError:
units_difference = dt_difference / ("1" + freq)
return units_difference == gap_difference
return (to_offset(freq) * gap_difference) + last_training_date == first_testing_date


_holdout_validation_result = namedtuple(
Expand Down