summary.glmulti: Handling glmulti objects

Description Usage Arguments Details Value Author(s) References See Also

View source: R/glmulti.r

Description

These standard S3 functions can be applied on glmulti to print a short report, obtain a more detailed summary, or produce different types of graphics.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
# S3 summary method for class 'glmulti'
## S3 method for class 'glmulti'
summary(object, ...)

# S3 print method for class 'glmulti'
## S3 method for class 'glmulti'
print(x, ...)


# S3 plot method for class 'glmulti'
## S3 method for class 'glmulti'
plot(x, type="p", ...)

Arguments

object

A glmulti object

x

An object of class glmulti

type

The type of graph to be produced. One of "p", "r", "s" or "w" (see below).

...

Further arguments.

Details

The name of the returned object is the name of the first object passed, with "consensus-" prepended. The params slot of the returned object is taken from the first object too.

Value

plot can be used to have a graphical representation of the results. Two types are proposed:

type="p" plots the IC profile (the IC values form the best to the worst model). A horizontal line delineates models that are less than 2 IC units away from the best model.

type="r" shows diagnostics of the fit (residuals versus predicted values, and QQ plot of residuals) for the five best models. It calls the plot functions on the fitted model objects, which should be defined (e.g. plot.lm, plot.glm). This type of plot can only be used if model objects are included in the glmulti object (i.e. if includeobjects was set to true).

type="s" plots the relative importance of model terms, i.e. the overall support for each variable across all models. A vertical line is drawn at 80

type="w" plots the normalized evidence weights of the models. A vertical line delineates models that sum um to 95

print prints a brief synthesis of the analysis (e.g. the best model found, its IC value and evidence weight...)

summary returns a list with more detailed elements:

name

the name of the analysis

method

The method used

fitting

The fitting function used

crit

The IC used

level

Whether interactions between predictors were considered or not

marginality

Whether the marginality rule was applied

confsetsize

The requested size of the confidence set

bestic

The lowest IC found

icvalues

The IC values of the models in the confidence set, from lower to greater

bestmodel

A list containing the formula of the best model found, or of the best models found if several had the same IC value

generations

The number of generations it took to converge. For genetic algorithm only.

elapsed

The actual (system) time it took. For genetic algorithm only.

includeobjects

A boolean indicating whether fitted model objects are contained in the object.

Author(s)

Vincent Calcagno, McGill University

References

Calcagno \& de Mazancourt 2010 J. Stat. Soft. v34 i12. See http://www.jstatsoft.org/v34/i12

See Also

glmulti


glmulti documentation built on July 2, 2020, 3:19 a.m.