| Multiplicative interaction {gnm} | R Documentation |
Mult is a symbolic wrapper to indicate a multiplicative interaction in
the formula argument to gnm.
Exp is a symbolic wrapper to indicate an exponeniated multiplier
in a Mult term.
Mult(..., multiplicity = 1) Exp(multiplier)
... |
a comma-separated list of two or more symbolic expressions, which represent the constituent multipliers in each component of the interaction. |
multiplicity |
a positive integer specifying the number of components to fit. |
multiplier |
a symbolic expression. |
A multiplicative interaction is either (if multiplicity = 1)
a product of the form
m_1 m_2 ... m_n,
where the constituent multipliers m_1, m_2, ..., m_n are linear
predictors; or (if multiplicity is greater than 1) a sum of
such “components”, each of which is a product of the same
linear predictors (but with different parameters).
Mult may be used to specify a multiplicative interaction in the
formula argument to gnm. The desired number of
components in the interaction is specified by multiplicity.
Models for the constituent multipliers are specified symbolically
as unspecified arguments to Mult. These symbolic expressions
are interpreted in the same way as the right hand side of a formula in
an object of class "formula". In particular, an intercept term
is added by default. Offsets can be added to constituent multipliers,
using offset.
Exp is a companion function to Mult, which may be used to
specify an exponentiated constituent multiplier, so that the
multiplier is constrained to be nonnegative.
Mult returns a list of objects of class "Mult", with one element
for each multiplicative component. A "Mult" object is a list of deparsed
expressions representing the factors in the multiplicative interaction.
Exp returns an object of class "Exp" which is a deparsed
expression representing the multiplier to be exponentiated.
These functions should only be used in formulae passed to gnm.
Heather Turner
Vargas, M, Crossa, J, van Eeuwijk, F, Sayre, K D and Reynolds, M P (2001). Interpreting treatment by environment interaction in agronomy trials. Agronomy Journal 93, 949–960.
set.seed(1)
## Using 'Mult' with 'Exp' to constrain the first constituent multiplier
data(yaish)
## Fit the "UNIDIFF" mobility model across education levels
unidiff <- gnm(Freq ~ educ*orig + educ*dest +
Mult(Exp(-1 + educ), -1 + orig:dest),
family = poisson, data = yaish, subset = (dest != 7))
## Not run:
## (this example can take quite a while to run)
##
## Using 'multiplicity' > 1
data(wheat)
yield.scaled <- wheat$yield * sqrt(3/1000)
treatment <- factor(paste(wheat$tillage, wheat$summerCrop, wheat$manure,
wheat$N, sep = ""))
bilinear2 <- gnm(yield.scaled ~ year + treatment +
Mult(year, treatment, multiplicity = 2),
family = gaussian, data = wheat)
## End(Not run)