Talk:What is ρ in the Bivariate Normal distribution?

From ShotStat
Jump to: navigation, search

It is customary to use \(\beta\) for the true slope parameter and \(b\) or \(\hat{\beta}\) for its ordinary least squares (OLS) estimate. Armadillo

Good point. The overall convention is to use Greek letters for true values and "standard" letters for sample values. I'll fix it.

Herb (talk) 12:04, 10 June 2015 (EDT)

Total least squares

It seems to me that this out to be done a bit differently that I outlined.

(1) calculate COI and translate axes to that point (2) determine a proportionality constant that related horizontal and vertical axes. The proportionality constant is determined so that errors in horizontal and vertical minimized.

given the above if B not statistically significant than zero then \(\rho\) would be zero. If b non-zero then \(\rho\) fraction of error accounted for by line.

Actually it seems more like this would be a second test.
Herb (talk) 19:58, 10 June 2015 (EDT)