influence.measures {base} | R Documentation |
This suite of functions can be used to compute some of the regression diagnostics discussed in Belsley, Kuh and Welsch (1980), and in Cook and Weisberg (1982).
influence.measures(lm.obj) summary.infl (object, digits = max(2, getOption("digits") - 5), ...) print.infl (x, digits = max(3, getOption("digits") - 4), ...) rstandard(lm.obj, infl = lm.influence(lm.obj), res = weighted.residuals(lm.obj), sd = sqrt(deviance(lm.obj)/df.residual(lm.obj))) rstudent (lm.obj, infl = ..., res = ...) dffits (lm.obj, infl = ..., res = ...) dfbetas (lm.obj, infl = ...) covratio (lm.obj, infl = ..., res = ...) cooks.distance(lm.obj, infl = ..., res = ..., sd = ...) hat(xmat)
lm.obj |
the resulting object returned by lm . |
infl |
influence structure as returned by lm.influence . |
res |
(possibly weighted) residuals, with proper default. |
sd |
standard deviation to use, see default. |
xmat |
the `X' or design matrix. |
The primary function is influence.measures
which produces a
class "infl"
object tabular display showing the DFBETAS for
each model variable, DFFITS, covariance ratios, Cook's distances and
the diagonal elements of the hat matrix. Cases which are influential
with respect to any of these measures are marked with an asterisk.
The functions dfbetas
, dffits
,
covratio
and cooks.distance
provide direct access to the
corresponding diagnostic quantities. Functions rstandard
and
rstudent
give the standardized and Studentized residuals
respectively. (These re-normalize the residuals to have unit variance,
using an overall and leave-one-out measure of the error variance
respectively.)
The optional infl
, res
and sd
arguments are there
to encourage the use of these direct access functions, in situations
where, e.g., the underlying basic influence measures (from
lm.influence
) are already available.
Note that cases with weights == 0
are dropped from all
these functions.
Belsley, D. A., Kuh, E. and Welsch, R. E. (1980) Regression Diagnostics. New York: Wiley.
Cook, R. D. and Weisber,g S. (1982) Residuals and Influence in Regression. London: Chapman and Hall.
## Analysis of the life-cycle savings data ## given in Belsley, Kuh and Welsch. data(LifeCycleSavings) lm.SR <- lm(sr ~ pop15 + pop75 + dpi + ddpi, data = LifeCycleSavings) summary(inflm.SR <- influence.measures(lm.SR)) inflm.SR which(apply(inflm.SR$is.inf, 1, any)) # which observations `are' influential dim(dfb <- dfbetas(lm.SR)) # the 1st columns of influence.measures all(dfb == inflm.SR$infmat[, 1:5]) rstandard(lm.SR) rstudent(lm.SR) dffits(lm.SR) covratio(lm.SR) ## Huber's data [Atkinson 1985] xh <- c(-4:0, 10) yh <- c(2.48, .73, -.04, -1.44, -1.32, 0) summary(lmH <- lm(yh ~ xh)) influence.measures(lmH)