hllm

fit.hllm

stepwise.hllm

hllm-class

an object of class <code><a rd-options="" href="/link/formula?package=gRbase&version=1.7-5" data-mini-rdoc="gRbase::formula">formula</a></code>. The right
    hand side of the formula is a list of the generators separated by
    <code>+</code>. A generator is specified by variable names with separated
    by <code>*</code>. Commonly used models have short hand notations:
    saturated model (<code>~.^.</code>), main effects (<code>~.^1</code>), all k'th
    order interactions (<code>~.^k</code>). 

formula

an object of class <code><a rd-options="" href="/link/gmData?package=gRbase&version=1.7-5" data-mini-rdoc="gRbase::gmData">gmData</a></code>.

gmData

an optional argument specifying a subset of the
    variables from the <code>gmData</code> object.

marginal

object

Definining the fitting engine. For hllm objects only
    "loglm" is implemented.

engine


  NOTICE: THIS FUNCTION IS DEFUNCT. PLEASE USE THE gRim PACKAGE FOR
  HIERARCHICAL LOG LINEAR MODELS.
  
  An implementation of hierarchical log-linear models using the
  framework of <code><a rd-options="" href="/link/gRbase?package=gRbase&version=1.7-5" data-mini-rdoc="gRbase::gRbase">gRbase</a></code>. A model object is defined using
  <code>hllm</code>, fitted using <code>fit</code> (which calls <code><a rd-options="MASS" href="/link/loglm?package=gRbase&version=1.7-5&to=MASS" data-mini-rdoc="MASS::loglm">loglm</a></code>)
  and a model search performed 
  using <code>stepwise</code>. The models may be displayed and manipulated
  using the <code><a rd-options="" href="/link/gRbase?package=gRbase&version=1.7-5" data-mini-rdoc="gRbase::gRbase">gRbase</a></code>.


models

internal

zzz-hllm

The 'gRbase' package provides general features
which are used by other graphical modelling packages, in particular
by the packages 'gRain', 'gRim' and 'gRc'.

'gRbase' contains several datasets relevant for use in connection with
graphical models. Almost all datasets used in the book Graphical
Models with R (2012) are contained in 'gRbase'.

'gRbase' implements several graph algorithms (based mainly on
representing graphs as adjacency matrices - either in the form
of a standard matrix or a sparse matrix). Some graph
algorithms are:
(i) maximum cardinality search (for marked and unmarked graphs).
(ii) moralize.
(iii) triangulate.
(iv) junction tree.

'gRbase' facilitates array operations,

'gRbase' implements functions for testing for conditional independence.

'gRbase' illustrates how hierarchical log-linear models may be
implemented and describes concept of graphical meta
data. These features, however, are not maintained anymore and
remains in 'gRbase' only because there exists a paper describing
these facilities: A Common Platform for Graphical Models in R:
The 'gRbase' Package, Journal of Statistical Software, Vol 14,
No 17, 2005.

NOTICE Proper functionality of 'gRbase' requires that the packages graph,
'Rgraphviz' and 'RBGL' are installed from 'bioconductor'; for
installation instructions please refer to the webpage given below.

zzz-hllm: Old: Hierarchical log-linear models

Description

Usage

Arguments

Value

See Also