Loading content
Concepts and practices for transforming raw data into informative features to improve predictive models.
Feature engineering is the process of transforming raw data into informative features that improve model generalization. It includes selection, creation, scaling and encoding of features as well as domain knowledge to boost predictive performance. Properly applied it reduces model complexity and improves interpretability.
Measure change of metrics (e.g. AUC, RMSE) after introducing new features.
Count of active features in production feed to control complexity.
Frequency of significant distribution changes in features in production.
Aggregated sums, means and counts per customer over defined time windows to predict purchase behavior.
Target encoding for high-cardinality categories with regularization to reduce overfitting.
Derive features such as trend, seasonality and time-based aggregates from event logs.
Data exploration and hypothesis formation.
Create and validate prototype features locally.
Automate recurring transformations in pipelines.
Version and document feature definitions.
Implement monitoring and define drift actions.