[enhancement] Add conditional seasonality as a new feature #1067
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
*Indicates whether conditional seasonality is applied or not - Type: boolean
*Mask out all the seasonality features that do not belong to the conditioned seasonality *Ensured conditional seasonality is part of the model's data_params *Integrated check in check_dataframe to rule out false column names for conditional seasonality
judussoari
added
enhancement
status: needs review
Model Benchmark
Model training plotsModel TrainingPeytonManningYosemiteTempsAirPassengers |
Codecov Report
@@ Coverage Diff @@
## main #1067 +/- ##
==========================================
+ Coverage 89.98% 90.01% +0.02%
==========================================
Files 21 21
Lines 4784 4805 +21
==========================================
+ Hits 4305 4325 +20
- Misses 479 480 +1
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
judussoari
requested review from
ourownstory and
Kevin-Chen0
and removed request for
ourownstory and
Kevin-Chen0
ourownstory
reviewed
Dec 21, 2022
…nditional_seasonality
ourownstory
requested changes
Jan 13, 2023
| If condition_name is provided, the dataframe passed to `fit` and | ||
| `predict` should have a column with the specified condition_name | ||
| containing only zeros and ones, deciding when to apply seasonality. | ||
| Floats between 0 and 1 can be used to apply seasonality partially. |
There was a problem hiding this comment.
Can you please add a (code) example - using quarterly changing weekly seasonality - to the docstring?
ourownstory
approved these changes
Jan 13, 2023
3 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
🔬 Background
NeuralProphet forces a weekly, daily, and yearly seasonality to the data (if activated). Apart from that, we were able to add custom seasonalities to the model. However, both custom and default seasonalities are applied to the entire dataset, without the possibility to specify a condition for a custom seasonality.
🔮 Key changes
condition_nameis a new input parameter form.add_seasonality(). Using this, the NP model fits and predicts seasonalities according to the condition specified by the user.The condition is specified through a new column that is added to the input
df. The new column should exclusively contain boolean values.True, if the condition applies to the row timestamp, andFalseotherwise. The same column(s) must be added to thefuturedataframe, if we want to predict into the unknown future.The user workflow follows the example of Facebook Prophet
See
test_custom_seasons()for an exmaple: on-season and off-season weekly seasonalities for Football. on-season we see a high during the weekends, while during the week there is no game going on. off-season does not show an increase during the weekend.📋 Review Checklist