A feature is a characteristic extracted that has been Prepared* for our ML Model to ingest to infer prediction
A feature is an input signal that carries information useful for a task. Concrete:
- Image → pixel intensity, edge, texture
- Text → word presence, embedding dimension
- Audio → frequency energy
- Tabular → age, income, clicks In neural nets:
- Early layers → simple features
- Deeper layers → abstract features
- Each neuron learns to detect one pattern (a feature)