modelTable: Collect results for different regression models

In regr: Building Regression Models

Description

modelTable produces a table that shows which variables are contained in various models. Typically, the different models are the results of various model selection functions and model development strategies.

format.modelTable and print.modelTable are used to turn the results into a nice form.

compareTerms returns a table indicating which terms appear in which models.

Usage

modelTable(models, seq = NULL) 

## S3 method for class 'modelTable'
format(x, digits=i.getoption("digits"),
       stars = c("***","** ","*  ",":  ",".  "), sep="", ...)
## S3 method for class 'modelTable'
print(x, tex = FALSE, transpose = FALSE, ...)

compareTerms(... , list=NULL, seq = NULL)

Arguments

models

a character vector containing the names of the models to be collected, or a list of model fits.

seq	The sequence in which the terms should appear in the table. If incomplete, the terms mentioned in seq will be the first rows in the result.
x	object of class modelTable
digits	to be used in formatting
stars	symbols used to indicate level of significance. They will be used for the intervals delimited by c(-1, 0.001, 0.01, 0.05, 0.1, 1)
sep	separator. "\&" is used to prepare LaTeX format.
tex	if TRUE, the output can be used in (La-)TeX.
transpose	logical. If TRUE, the table is printed in transposed form
...	other arguments, passed to print or format. In compareTerms, the models for which the terms should be compared can be listed as in c or list etc.
list	A list of models.

Value

modelTable produces a list of class "modelTable", with components

coef	the coefficients of the models, collected in a matrix. The columns of the matrix correspond to the models, the rows, to the terms. For terms with more than 1 degree of freedom, the matrix element is NA. If possible, standardized coefficients are given.
p	p values, in the same form

sigma	estimated error standard deviations of the models
fitfun	fitting functions used for fitting the models

format.modelTable returns a character matrix and thereby prepares for print.modelTable. If needed, the ouput of format.modelTable can be reordered or subsetted.

Author(s)

Werner A. Stahel, ETH Zurich

See Also

on formatting and printing of P-values, the standard R functions symnum, format.pval.

Examples

data(d.blast, package="plgraphics")
r.blast <-
  regr(log10(tremor)~location+log10(distance)+log10(charge), data=d.blast)
r.bl2 <- update(r.blast, ~.+location:log10(distance)+I(log10(charge)^2) )
r.mt <- modelTable(list(large=r.bl2, original=r.blast))
r.mt

format(r.mt)
print(r.mt, tex=TRUE)
compareTerms(r.blast, r.bl2, seq=c("log10(distance)", "log10(charge)"))

regr documentation built on Oct. 22, 2019, 3:01 p.m.