Feature extraction Machine Learning
noun phrase
Definition: The process of transforming raw data into informative features for machine learning; feature extraction often refers to deriving new representations, while feature engineering is a broader term that also includes selecting, creating, and transforming features [scikit-learn documentation; IBM].
Example in context: “These networks can directly process raw data freeing researchers from time-consuming manual feature extraction.” [Mattia et al. 2024]
“Moreover, the edge layer also performs feature engineering and preprocessing of sensor data.” [Bala et al. 2024]
Synonym: feature engineering (broader in many contexts)
Related terms: representation learning, feature construction, feature transformation