---
library_name: autogluon
license: mit
tags:
- tabular-classification
- automl
- autogluon
datasets:
- Iris314/Youtube_music_comments
metrics:
- accuracy
- f1
---

# Tabular AutoML (AutoGluon)

**Task:** Classification
**Dataset:** `Iris314/Youtube_music_comments` (classmate HW1 dataset)
**Target column:** `label`

## Best Model
- **Name:** `WeightedEnsemble_L2`
- **Hyperparameters:** See below (AutoGluon model_hyperparams for best model)

## Results
- **Hold-out (same distribution as training):**
  - Accuracy = 0.9905
  - Weighted F1 = 0.9905

- **Additional check (original split):**
  - Accuracy = 1.0000
  - Weighted F1 = 1.0000

## Notes
- Trained with `presets="best_quality"`, `time_limit=300`.
- AutoGluon ensembles classical ML models and tunes hyperparameters automatically.