Add conformal prediction intervals support for Hyper-Tree models - #9
Merged
Conversation
StatMixedML
commented
Jun 3, 2026
StatMixedML
commented
Owner
- Introduced ForecastIntervals class for configuring and calibrating prediction intervals.
- Updated HyperTreeAR, HyperTreeETS, and HyperTreeNetAR to support conformal intervals.
- Enhanced forecast method to include interval columns for specified confidence levels.
- Added unit tests for conformal prediction functionality.
- Updated README to reflect new features and usage examples.
- Introduced ForecastIntervals class for configuring and calibrating prediction intervals. - Updated HyperTreeAR, HyperTreeETS, and HyperTreeNetAR to support conformal intervals. - Enhanced forecast method to include interval columns for specified confidence levels. - Added unit tests for conformal prediction functionality. - Updated README to reflect new features and usage examples.
…tion for compatibility
…er window Revert CI torch install to --extra-index-url (matches passing main)
…nd interval_columns
|
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #9 +/- ##
==========================================
+ Coverage 93.37% 95.17% +1.80%
==========================================
Files 7 15 +8
Lines 1207 3462 +2255
Branches 207 623 +416
==========================================
+ Hits 1127 3295 +2168
- Misses 42 78 +36
- Partials 38 89 +51 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
StatMixedML
marked this pull request as draft
June 3, 2026 13:40
…Nixtla attributions
- Fix HyperTreeNetAR training a hidden model copy (lgb.train params deepcopy):
new NoDeepcopyObjective binds objectives to the live instance, replacing the
class-level network state that leaked weights across instances
- HyperTreeAR: add hessian_method="analytic" (new default), closed-form
grad/Hessian via the AR fit's linearity
- HyperTreeETS: add seasonal_init="classical"|"legacy" (statsforecast-style
init vs. verbatim pre-0.2.0; experiments default to legacy), cache the init
per dataset, validate seasonal positions
- HyperTreeSTL: make the trend-smoothing window learnable, fix short-horizon
crashes
- Reject nn.L1Loss in all models (zero curvature breaks Newton boosting)
- Mask-aware conformal residuals; reject duplicate dates; docs and ~30 tests
- HyperTreeVAR / HyperTreeNetVAR: VAR(p) over aligned panels with
per-series scaling and a restricted GVAR-style factor design
(type="factor")
- HyperTreeTSB: intermittent demand with feature-driven smoothing rates
- Additive-seasonality option for HyperTreeETS (ets_type="additive")
- Example notebooks, README table rows, and release notes
- HyperTreeVAR / HyperTreeNetVAR: VAR(p) over aligned panels with
per-series scaling and a restricted GVAR-style factor design
(type="factor")
- HyperTreeTSB: intermittent demand with feature-driven smoothing rates
- Additive-seasonality option for HyperTreeETS (ets_type="additive")
- Example notebooks, README table rows, and release notes
- HyperTreeVAR / HyperTreeNetVAR: VAR(p) over aligned panels with
per-series scaling and a restricted GVAR-style factor design
(type="factor")
- HyperTreeTSB: intermittent demand with feature-driven smoothing rates
- Additive-seasonality option for HyperTreeETS (ets_type="additive")
- Example notebooks, README table rows, and release notes
…Tree models; Adjust STL model for conformal intervals and Hessian stabilization
…o tag and clarify reproducibility requirements
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.