Find best default values for LGBM Classifier
Offen
enhancement
- Vorherrschende Sprache
- Python
- Sterne
- 850
- Forks
- 96
- PR-Merge-Kennzahlen
- Keine gemergten PRs in 30 T.
Beschreibung
Like the LGBM Regressor perf tests [here](https://alteryx.quip.com/HTTuArJve1uW/LightGBM-Regressor-Perf-Tests), this issue tracks creating perf tests to find the best default parameters for `n_estimators` and `max_depth` for LGBM Classifier, rather than using the defaults that LightGBM uses.
For our regressor component, we see that we achieve both speed and performance gains on by changing these values, so we should check for similar patterns in the classifier.
Beitragsleitfaden
Bewertung
Dieses Issue wurde noch nicht bewertet.