# Feature Scaling

> Puts number columns on one range. Then big numbers do not drown out small ones.

**Field:** Artificial Intelligence  
**Category:** Machine Learning  
**Also known as:** Normalization (Machine Learning), Standardization (Machine Learning), Min-Max Scaling

## Think of it like

It is like putting grams and tons in one unit. Then you can compare them.

## Definition

Feature scaling puts number columns on a similar range before training. One column might be age, from 0 to 100. Another might be income, up to 200,000. Without scaling, the big numbers can drown out the small ones. Min-max scaling squeezes each column to 0 to 1. Standardization shifts it so the average is 0 and most values sit near it.

## Example

Emma's model reads age and income. After scaling, both run from 0 to 1. Her k-nearest neighbors model stops ignoring age.

## Why it matters

Models that measure distance or use gradient descent learn badly without it. Tree models do not need it.

## Common mistake

Scaling with numbers taken from the whole dataset. Fit the scaler on the training set only, or test data leaks in.

## Three principles

- **Same range.** Columns become easy to compare.
- **Two common ways.** Squeeze to 0 to 1, or center on 0.
- **Fit on training.** Then apply it to the rest.

## Related terms

- [Feature (Machine Learning)](https://learn.benibix.com/term/ai/feature)
- [Feature Engineering](https://learn.benibix.com/term/ai/feature-engineering)
- [k-Nearest Neighbors (k-NN)](https://learn.benibix.com/term/ai/k-nearest-neighbors)
- [Gradient Descent](https://learn.benibix.com/term/ai/gradient-descent)
- [Data Leakage](https://learn.benibix.com/term/ai/data-leakage)

## Told apart from

- [Feature Scaling vs Database Normalization](https://learn.benibix.com/compare/feature-scaling-vs-database-normalization): People hear normalization and think of one idea, and there are two. In machine learning, it means scaling number columns to a similar range. In databases, it means splitting tables so each fact is stored once. One changes the numbers. The other changes the table layout.

## Taught in

- [Feature Engineering](https://learn.benibix.com/learn/ai-machine-learning/feature-engineering): Machine Learning, 4/18

![A raw row with city Paris, size 50 square metres and price 600 becomes a ready row. City turns into three columns, London 0, Paris 1, Tokyo 0, marked one-hot. A new column, price per square metre, reads 12. Price is scaled to 0.67. A label column reading sold fast is added by a person.](https://learn.benibix.com/diagrams/ai3-data-prep.svg)

---

Source: https://learn.benibix.com/term/ai/feature-scaling
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-17
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
