Why do we square the errors in OLS? Why not use true values?
| Jared Baumeister | 08/26/26 | | Karlstack | 08/26/26 | | Jared Baumeister | 08/26/26 | | coalition of rabbis | 08/26/26 | | Jared Baumeister | 08/26/26 | | Jared Baumeister | 08/26/26 | | Jared Baumeister | 08/26/26 |
Poast new message in this thread
Date: August 26th, 2026 12:17 AM Author: Jared Baumeister
Squaring has a real cost. Because a big error is punished so hard, a single wrong data point can drag the whole line towards itself. One student recorded as 8 hours and 3 marks, by a typo, will bend the line more than the other seven points together.
The absolute version is far more resistant to that, and it has a name, mean absolute error. It is a good choice when your data has outliers you cannot remove. It is used less often mostly because of the corner, and because squared error has clean mathematics behind it.
https://teachyourselfcoding.com/courses/learn-machine-learning-from-first-principles-using-javascript/lesson/why-we-square-the-error/?page_tab=overview
(http://www.autoadmit.com/thread.php?thread_id=5897341&forum_id=2),#50093513) |
|
|