MATIH Platform is in active MVP development. Documentation reflects current implementation status.
13. ML Service & MLOps
AutoML
Feature Engineering

Automated Feature Engineering

The AutoML feature engineering module automatically discovers and applies feature transformations to improve model performance.


Transformations

CategoryTransforms
NumericStandardization, normalization, log transform, polynomial features
CategoricalOne-hot encoding, target encoding, ordinal encoding
TemporalDate parts, cyclical encoding, lag features, rolling statistics
TextTF-IDF, word embeddings, sentiment scores
InteractionPairwise products, ratios, differences

Source Files

FilePath
Feature Engineeringdata-plane/ml-service/src/automl/feature_engineering.py