Prediction intervals example noteook #189
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
🚀 Deployed on https://deploy-preview-189--etna-docs.netlify.app |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #189 +/- ##
=======================================
Coverage 89.29% 89.29%
=======================================
Files 198 198
Lines 12866 12866
=======================================
Hits 11489 11489
Misses 1377 1377 ☔ View full report in Codecov by Sentry. |
| @@ -15,7 +15,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 | |||
| - `ConformalPredictionIntervals` method for prediction intervals estimation ([#152](https://github.com/etna-team/etna/pull/152)) | |||
| - Add DeepARNativeModel ([#114](https://github.com/etna-team/etna/pull/114)) | |||
| - `EmpiricalPredictionIntervals` method for prediction intervals estimation ([#173](https://github.com/etna-team/etna/pull/173)) | |||
| - | |||
| - Prediction intervals tutorial notebook ([#189](https://github.com/etna-team/etna/pull/189)) | |||
There was a problem hiding this comment.
We should also prepare this notebook for a tutorials documentation page. See changelog with title "Adding a new jupyter notebook tutorial".
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:50Z 1). We should probably describe what prediction interval is and may be there to read about it.
2). We should provide a distinction between model-specific methods that are implemented in our models and model-agnostic methods which are implemented separately.
3). I don't think that current description is clear enough with this quantiles vs. arbitrary borders. |
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:51Z Line #2. forecast This produces a very big output. Could we for example, make it smaller by using |
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:52Z May be it is better to say "in the dataset with forecast"?
|
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:53Z I don't think the second sentence is necessary and could be removed. |
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:53Z There are should be no intervals in the dataset
It doesn't look good. Try using https://www.deepl.com/write
Interval border sets should match for all the segments. I don't understand what it means.
|
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:54Z This method estimate estimates |
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:55Z ConstantWidthInterval isn't rendered well here at reviewnb. |
|
View / edit / reply to this conversation on ReviewNB d-a-bunin commented on 2023-12-13T10:29:56Z Why don't we have any conclusion? |
Before submitting (must do checklist)
Proposed Changes
Closing issues
closes #170