What is a feature in machine learning?

Features are the language models speak.
Thomas Wolf

How It Works:

A feature is an individual measurable property (e.g., age, transaction amount) used as input for training models.

Key Benefits:

  • Signal clarity: Good features boost predictive power.
  • Model simplicity: Well-engineered features reduce complexity.
  • Insightful analysis: Features reveal data patterns.

Real-World Use Cases:

  • Customer churn: Features like session frequency.
  • Spam detection: Features like email word counts.

FAQs

What's feature engineering?
How many features needed?