In linear regression, the way root mean square is calculated is by:
- Taking the distance of the point from the line
- Squaring the values (to get positive values)
- Taking the mean
- And taking square root
Why absolute values are not taken (which returns positive values) instead of squaring the values and then taking square root?