Lines Matching +refs:model +refs:matrix +refs:lm

4 % 2009-09-16: added argument singular.ok to lm method. J. Fox
5 % 2012-04-04: weighted lm now allowed. John
12 \alias{hccm.lm}
23 hccm(model, ...)
25 \method{hccm}{lm}(model, type=c("hc3", "hc0", "hc1", "hc2", "hc4"), 
28 \method{hccm}{default}(model, ...)
32   \item{model}{a unweighted or weighted linear model, produced by \code{lm}.}
36   \item{singular.ok}{if \code{FALSE} (the default is \code{TRUE}), a model with aliased coefficients
38     matrix that's returned.}
39   \item{...}{arguments to pass to \code{hccm.lm}.}
43 …The original White-corrected coefficient covariance matrix (\code{"hc0"}) for an unweighted model …
45   where \eqn{e_{i}^{2}}{e^2} are the squared residuals, and \eqn{X} is the model
46 …matrix. The other methods represent adjustments to this formula. If there are weights, these are i…
47   corrected covariance matrix.
49   The function \code{hccm.default} simply catches non-\code{lm} objects.
54   The heteroscedasticity-corrected covariance matrix for the model.
73   Using heteroscedasity consistent standard errors in the linear regression model. 
77   A heteroskedastic consistent covariance matrix estimator and a direct test of heteroskedasticity.
84 mod<-lm(interlocks~assets+nation, data=Ornstein)