Help System (web edition)

Linear regression analysis
The linear regression model is specified by activating variables in the
data set as follows:
  Y is the regressand,
  X is a regressor,
  R is the residual variable (optional, must not be protected),
  P is the variable for predicted values (optional, not protected).
For general rules of variable selection, see MASK and VARS.

The following operations work on this basis:
  A = LINREG  (OLS through correlations) 
  B = REGDIAG (OLS by orthogonalization, regression diagnostics) 
  C = MOVREG  (Moving linear regression analysis) (by KV) 
  D = ROBREG  (Least Median of Squares regression analysis) (by Reino Siren) 
  E = ESTIMATE (for non-linear models) 

  U = Sucros for regression analysis (special applications) 
  S = More information on statistical analysis