Min-Max Scaling

Topics: Data Processing

Concept

Min-Max scaling is a method used to normalize measurements that have a definite minimum and maximum into the range into a defined fixed range. The process maps the minimum to $0$ and the maximum to $1$, resulting in a normalized range of $[0,1]$. From there, the range can be changed using a gain and an offset.

Let $r$ be the raw measurement. Then the normalized value $n$ in the range $[0,1]$ is as follows:

$$n = \frac{r - min(r)}{max(r) - min(r)}$$

To modify the range into any desired range:

$$n' = \alpha \cdot n + \beta$$

where $\alpha$ is a gain and $\beta$ is an offset.

References