Feature Vector

Category: science

A row of numerical data representing all the features for a single data point.

The vector is the "input line." It is what the model actually "sees." If you have 50 features for one customer (Age, Credit Score, etc.), the vector is an array of 50 numbers. This array is the mathematical input that drives every prediction.

Common Examples

• The software generates a feature vector in real-time for every new claim, which is then fed into our production scoring engine.
• Standardizing the shape and order of the feature vector across all data pipelines is critical for consistent model execution performance.