Loglinear Model Utilities

These functions generate lists of terms to specify a loglinear model in a form compatible with loglin and also provide for conversion to an equivalent loglm specification or a shorthand character string representation.

They allow for a more conceptual way to specify such models by a function for their type, as opposed to just an uninterpreted list of model terms and also allow easy specification of marginal models for a given contingency table.

They are intended to be used as tools in higher-level modeling and graphics functions, but can also be used directly.

Usage

conditional(nf, table = NULL, factors = 1:nf, with = nf)

joint(nf, table = NULL, factors = 1:nf, with = nf)

markov(nf, factors = 1:nf, order = 1)

mutual(nf, table = NULL, factors = 1:nf)

saturated(nf, table = NULL, factors = 1:nf)

loglin2formula(x, env = parent.frame())

loglin2string(x, brackets = c("[", "]"), sep = ",", collapse = " ", abbrev)

Arguments

nf

number of factors for which to generate the model

table

a contingency table used only for factor names in the model, typically the output from table and possibly permuted with aperm

factors

names of factors used in the model formula when table is not specified

with

For joint and conditional models, with gives the indices of the factors against which all others are considered jointly or conditionally independent

order

For markov, this gives the order of the Markov chain model for the factors. An order=1 Markov chain allows associations among sequential pairs of factors, e.g., [A,B], [B,C], [C,D] .... An order=2 Markov chain allows associations among sequential triples.

x

For the loglin2* functions, a list of terms in a loglinear model, such as returned by conditional, joint, ...

env

For loglin2formula, environment in which to evaluate the formula

brackets

For loglin2string, characters to use to surround model terms. Either a single character string containing two characters (e.g., '[]' or a character vector of length two.

sep

For loglin2string, the separator character string used for factor names within a given model term

collapse

For loglin2string, the character string used between terms in the the model string

abbrev

For loglin2string, whether and how to abbreviate the terms in the string representation. This has not yet been implemented.

Details

The main model specification functions, conditional, joint, markov, ..., saturated, return a list of vectors indicating the marginal totals to be fit, via the margin argument to loglin. Each element of this list corresponds to a high-order term in a hierarchical loglinear model, where, e.g., a term like c("A", "B") is equivalent to the loglm term "A:B" and hence automatically includes all low-order terms.

Note that these can be used to supply the expected argument for the default mosaic function, when the data is supplied as a contingency table.

The table below shows some typical results in terms of the standard shorthand notation for loglinear models, with factors A, B, C, ..., where brackets are used to delimit the high-order terms in the loglinear model.

function	3-way	4-way	5-way
mutual	[A] [B] [C]	[A] [B] [C] [D]	[A] [B] [C] [D] [E]
joint	[AB] [C]	[ABC] [D]	[ABCE] [E]
joint (with=1)	[A] [BC]	[A] [BCD]	[A] [BCDE]
conditional	[AC] [BC]	[AD] [BD] [CD]	[AE] [BE] [CE] [DE]
condit (with=1)	[AB] [AC]	[AB] [AC] [AD]	[AB] [AC] [AD] [AE]
markov (order=1)	[AB] [BC]	[AB] [BC] [CD]	[AB] [BC] [CD] [DE]
markov (order=2)	[A] [B] [C]	[ABC] [BCD]	[ABC] [BCD] [CDE]
saturated	[ABC]	[ABCD]	[ABCDE]

loglin2formula converts the output of one of these to a model formula suitable as the formula for of loglm.

loglin2string converts the output of one of these to a string describing the loglinear model in the shorthand bracket notation, e.g., "[A,B] [A,C]".

Value

For the main model specification functions, conditional, joint, markov, ..., the result is a list of vectors (terms), where the elements in each vector are the names of the factors. The elements of the list are given names term1, term2, ....

References

These functions were inspired by the original SAS implementation of mosaic displays, described in the User's Guide, http://www.datavis.ca/mosaics/mosaics.pdf

Author

Michael Friendly

See also

loglin, loglm

Examples

joint(3, table=HairEyeColor)
#> $term1
#> [1] "Hair" "Eye" 
#> 
#> $term2
#> [1] "Sex"
#> 
# as a formula or string
loglin2formula(joint(3, table=HairEyeColor))
#> ~Hair:Eye + Sex
#> <environment: 0x0000016cc610bcb8>
loglin2string(joint(3, table=HairEyeColor))
#> [1] "[Hair,Eye] [Sex]"

joint(2, HairEyeColor)  # marginal model for [Hair] [Eye]
#> $term1
#> [1] "Hair"
#> 
#> $term2
#> [1] "Eye"
#> 

# other possibilities
joint(4, factors=letters, with=1)
#> $term1
#> [1] "b" "c" "d"
#> 
#> $term2
#> [1] "a"
#> 
joint(5, factors=LETTERS)
#> $term1
#> [1] "A" "B" "C" "D"
#> 
#> $term2
#> [1] "E"
#> 
joint(5, factors=LETTERS, with=4:5)
#> $term1
#> [1] "A" "B" "C"
#> 
#> $term2
#> [1] "D" "E"
#> 

conditional(4)
#> $term1
#> [1] 1 4
#> 
#> $term2
#> [1] 2 4
#> 
#> $term3
#> [1] 3 4
#> 
conditional(4, with=3:4)
#> $term1
#> [1] 1 3 4
#> 
#> $term2
#> [1] 2 3 4
#> 

# use in mosaic displays or other strucplots
mosaic(HairEyeColor, expected=joint(3))

mosaic(HairEyeColor, expected=conditional(3))


# use with MASS::loglm
cond3 <- loglin2formula(conditional(3, table=HairEyeColor))
cond3 <- loglin2formula(conditional(3))  # same, with factors 1,2,3
require(MASS)
loglm(cond3, data=HairEyeColor)
#> Call:
#> loglm(formula = cond3, data = HairEyeColor)
#> 
#> Statistics:
#>                       X^2 df P(> X^2)
#> Likelihood Ratio 156.6779 18        0
#> Pearson          147.9440 18        0

saturated(3, HairEyeColor)
#> $term1
#> [1] "Hair" "Eye"  "Sex" 
#> 
loglin2formula(saturated(3, HairEyeColor))
#> ~Hair:Eye:Sex
#> <environment: 0x0000016cc610bcb8>
loglin2string(saturated(3, HairEyeColor))
#> [1] "[Hair,Eye,Sex]"
loglin2string(saturated(3, HairEyeColor), brackets='{}', sep=', ')
#> [1] "{Hair, Eye, Sex}"