The existing normalization / scaling methods use multiple samples of features for techniques like min-max normalization, or scaling by subtracting mean and dividing by standard deviation.
For example,
| speed (m/s) | acceleration (m/s^2) | |
|---|---|---|
| Vehicle 1 | 13.9 | 2.5 |
| Vehicle 2 | 7.5 | 4.2 |
| . | . | . |
| . | . | . |
| . | . | . |
| Vehicle n | 3.2 | -1.6 |
If I only had the value from vehicle 1 - 13.9 m/s and 2.5 m/s^2, how would I scale these values so that they are at comparable levels?