PUBDEV-6422 Add Python API support for integration of target encoding into AutoML #3760
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
PUBDEV-6319: Added Andrey Spiridonov's groupby test. PUBDEV-6319: Removed printOutFrameAsTable to reduce runtime.
…ge because we still loosing an original order of rows.
…ame to IcedHashMap.
…ng in h2o-automl module. Added test for BroadcastJoin (distributed case)
…g for a single node single chunk case.
…e. Added LeaveOneOut switch case.
…ass to reduce complexity of AutoML class and to reduce coupling by introducing clear interface/dependencies.
…dded FixedTEParamsStrategy to specify single target encoding params for whole automl process.
… test with comparison of encodings.
…strategy and fixed target encoding parameters strategy
…ed unit and integration tests.
…nto AutoML. Fix creation of the zero vecs with incompatible layouts in TE.
…Added StratifiedSampler.
…r evaluation during TE hyperparameters grid search.
…f target encoding.
…erboard frame. It is closer to our attempt to use sampling for hyper parameters search but seems that found parameters are not doing good during final scoring in automl.
… hp search. HP grid search's evaluations is performed based on the original data. Holdout.None strategy quite often hurts as we reduce amount of training data.
… sequence of models. Added benchmark that compares performance of thee first models with and without TE.
…ml with TE and automl without TE.
…equential Model-Based Optimization with DFR. Benchmarks for comparing results.
…or SMBO. Added theBiggerTheBetter parameter to support different metrics.
…t' into PUBDEV-5742_TE_automl_integration
…ne per tree predictions from DRF in order to get variances for predictions. It should go to a separate PR if we decide to stay with DRF as a surrogate model for SMBO." This reverts commit 67cbd03
… keys leakages in RGS TE params selection strategy.
…a tests for AutoMLTargetEncodingAssistant.
…gy. All related files were moved into a separate branch for future investigation.
…rch over categorical columns to apply target encoding to.
…. Fixed ignored tests in TargetEncodingHyperparamsEvaluatorTest.
…tionStrategy instead of multiple parameters.
… is no columns to encode. Added ability to easily switch off TE in AutoML
|
@deil87 please rebase & resolve conflicts |
|
Test comment - can i reopen this PR? |
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.
This PR was postponed until we are back on track for TE integration with AutoML. And we are not there yet. Full Mojo support comes first.