Preprocess data for machine learning and deep learning

You can use Databricks Feature Store to create new features, explore and re-use existing features, select features for training and scoring machine learning models, and publish features to low-latency online stores for real-time inference.

On large datasets, you can use Spark SQL and MLlib for feature engineering. Third-party libraries included in Databricks Runtime ML such as scikit-learn also provide useful helper methods. For examples, see the following machine learning notebooks for scikit-learn and MLlib:

• Feature engineering with scikit-learn
• Feature engineering with MLlib

For more complex deep learning feature processing, this example notebook illustrates how to use transfer learning for featurization:

• Featurization for transfer learning