Author: Michael Yee
Date: 08:56:38 04/06/04
Go up one level in this thread
On April 05, 2004 at 11:51:57, Roger D Davis wrote: >Tim, > >If you're using multivariate regression, doesn't that assume that the regression >variables are normally distributed? > >Roger I think (multiple) linear regression typically assumes that the *errors* (or disturbances) are normally distributed. It basically helps for computing certain things about the parameters (like t-stat/significance of parameters, variance of parameters, etc.). But you can minimize the sum of squares errors without bothering with distributions at all. It's just an optimization problem. Michael
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.