effects
Single Degree-of-freedom Effects from a Fitted Model
Description
Returns a vector or matrix of uncorrelated effects.
This function is generic; see the help file for Methods
for more information.
Method functions can be written to handle specific classes of data.
Classes that already have methods for this function
include "lm" and "glm".
Usage
effects(object, ...)
## S3 method for class 'lm':
effects(object, set.sign = FALSE, ...)
## S3 method for class 'glm':
effects(object, ...)
Arguments
object |
a fitted model, as returned by lm or a related
model fitting function.
|
set.sign |
a logical value; if TRUE,
the sign of the effects corresponding to coefficients in the model
will be set to agree with the signs of the corresponding coefficients.
Otherwise, the signs are determined by the algorithm for the orthogonal
decomposition to promote numerical stability.
|
Value
a vector or, if the response was a matrix, a matrix of effects.
These are single-degree-of-freedom values,
orthogonal and hence uncorrelated under standard linear model assumptions.
The first length(coef(object)) entries in the output will be labelled
with the coefficient names. For a least-squares problem their squares
are the sum of squares associated with that coefficient. The remaining
entries will not be labelled and sum of their squares is the residual
sum of squares.
If object inherits from the class lm, but not
also glm, the return value will also be given the
the class coef and the attribute assign, copied
from object.
Differences between TIBCO Enterprise Runtime for R and Open-source R
In open-source R, when effects component of object is NULL,
effects is stopped with an error. In TIBCO Enterprise Runtime for R, it returns NULL.
See Also
Examples
fit <- lm(data=ToothGrowth, len ~ supp + dose)
effects(fit)[1:3]^2
# compare to "Sum Sq" column in anova table
anova(fit)