[HN Gopher] Why does a least squares fit appear to have a bias w...
___________________________________________________________________
Why does a least squares fit appear to have a bias when applied to
simple data?
Author : azeemba
Score : 61 points
Date : 2026-01-04 20:25 UTC (2 hours ago)
(HTM) web link (stats.stackexchange.com)
(TXT) w3m dump (stats.stackexchange.com)
| charlieyu1 wrote:
| If you plot the regression line of y against x, and also x
| against y, you would get two different lines.
|
| I found it in the middle of teaching a stats class, and feel
| embarrassed.
|
| I guess normalising is one way to remove the bias.
| dllu wrote:
| You can think of it as: linear regression models only noise in y
| and not x, whereas ellipse/eigenvector of the PCA models noise in
| both x and y.
| analog31 wrote:
| That brings up an interesting issue, which is that many systems
| do have more noise in y than in x. For instance, time series
| data from an analog-to-digital converter, where time is based
| on a crystal oscillator.
| GardenLetter27 wrote:
| This fact underlies a lot of causal inference.
| sega_sai wrote:
| The least squares and pca minimize different loss functions. One
| is sum of squares of vertical(y) distances, another is is sum of
| closest distances to the line. That introduces the differences.
| ryang2718 wrote:
| I find it helpful to view least as fitting the noise to a
| Gaussian distribution.
| LudwigNagasena wrote:
| OLS estimator is the minimum-variance linear unbiased
| estimator even without the assumption of Gaussian
| distribution.
| gpcz wrote:
| You would probably get what you want with a Deming regression.
___________________________________________________________________
(page generated 2026-01-04 23:00 UTC)