Automated Feature Engineering
The AutoML feature engineering module automatically discovers and applies feature transformations to improve model performance.
Transformations
| Category | Transforms |
|---|---|
| Numeric | Standardization, normalization, log transform, polynomial features |
| Categorical | One-hot encoding, target encoding, ordinal encoding |
| Temporal | Date parts, cyclical encoding, lag features, rolling statistics |
| Text | TF-IDF, word embeddings, sentiment scores |
| Interaction | Pairwise products, ratios, differences |
Source Files
| File | Path |
|---|---|
| Feature Engineering | data-plane/ml-service/src/automl/feature_engineering.py |