2070 user specified date feature#2217
Conversation
…n to DatetimeFeaturizer and DelayedFeatureTransformer
Codecov Report
@@ Coverage Diff @@
## main #2217 +/- ##
=========================================
+ Coverage 100.0% 100.0% +0.1%
=========================================
Files 288 288
Lines 24449 24462 +13
=========================================
+ Hits 24431 24444 +13
Misses 18 18
Continue to review full report at Codecov.
|
# Conflicts: # evalml/automl/automl_search.py # evalml/tests/automl_tests/test_automl_search_classification.py # evalml/tests/automl_tests/test_automl_search_regression.py # evalml/tests/pipeline_tests/test_time_series_baseline_pipeline.py
ParthivNaresh
requested review from
dsherry and
freddyaboulton
and removed request for
dsherry and
freddyaboulton
ParthivNaresh
requested review from
angela97lin,
bchen1116,
chukarsten,
dsherry,
freddyaboulton and
jeremyliweishih
|
|
||
| on: | ||
| pull_request: | ||
| types: [ opened, synchronize ] |
There was a problem hiding this comment.
Do we want to keep this for feature builds moving forward?
There was a problem hiding this comment.
We can try it out and if it's too slow we can undo the change?
|
@ParthivNaresh I see this in the PR description:
This is no longer true, correct? If so please delete :) |
|
@dsherry Sorry I meant after this is merged, ARIMA will be updated in another PR :) |
freddyaboulton
left a comment
freddyaboulton
left a comment
There was a problem hiding this comment.
@ParthivNaresh Looks great!
|
|
||
| on: | ||
| pull_request: | ||
| types: [ opened, synchronize ] |
There was a problem hiding this comment.
We can try it out and if it's too slow we can undo the change?
| pd.testing.assert_index_equal(y_test.index, answer[i][1]) | ||
|
|
||
|
|
||
| @pytest.mark.parametrize("max_delay,gap,date_index", [(0, 0, "Date"), (1, 0, "Date"), (0, 1, "Date")]) |
There was a problem hiding this comment.
I think we can delete this test? The ts data splitter doesn't sort the data.
| pipeline = BinaryClassificationPipeline([MockEstimator]) | ||
| algo = IterativeAlgorithm(allowed_pipelines=[pipeline, pipeline, pipeline], | ||
| pipeline_params={'pipeline': {"gap": 2, "max_delay": 10}}, | ||
| pipeline_params={'pipeline': {'date_index': "Date", "gap": 2, "max_delay": 10}}, |
| ------------- | ||
| **Future Releases** | ||
| * Enhancements | ||
| * Added `date_index` as a required parameter for TimeSeries problems :pr:`2217` |
4 participants
Fixes #2070
This is a refiling of the old PR with the previous (main) version of ARIMA. After this is merged, ARIMA will be updated in a different PR using
sktimeas opposed tostatsmodels.