gmvalid-package {gmvalid}R Documentation

Validation of graphical models

Description

This package provides functions among others that can be used to analyse graphical models. This includes e.g. the possibility to simulate data sets given a dependence model, to analyze discrete graphical models utilizing the MIM program or the CoCo package and to quantify associations or interactions.

Furthermore, the uncertainty of a selected graphical model can be described using the bootstrap or the best prediction model can be evaluated for a dichotomous outcome variable and several discrete influences using cross validation.

Details

Package: gmvalid
Type: Package
Version: 1.2
Date: 2008-12-18
License: GPL (>=2)

  • This package requires the packages grid, epitools, gtools and CoCo.
  • The package mimR is optional as well as the corresponding software MIM that can be downloaded at http://www.hypergraph.dk.
  • The mimR package itself requires the package Rgraphviz that is found in the Bioconductor repository.
  • Further the packages grid, epitools and gtools are required.
  • Before using any functions that use MIM make sure that the program is runnning.
  • The package mimR will only work properly if your R working directory path does not contain hyphens ("-").
  • For mimR requirements we refer to the package's help page mimR and its homepage http://gbi.agrsci.dk/~shd/public/mimR/index.html.
  • If CoCo calculations fail, please type "endCoCo()" to remove temporary files.
  • CoCo crashes occasionally. If you have recurring problems with a specific data set, we suggest to use the MIM-option or to reboot the computer.
  • Model formulae have to be specified according to MIM model formulas or respectively CoCo model formulae.
  • To provide a package version that also runs on unix and therefore has no MIM functionalities, we had to put all MIM related examples in "Not Run". But for Windows the examples work.

    Note

    This work has been supported by the German Research Foundation
    (DFG: http://www.dfg.de) under grant scheme PI 345/2-1.

    Author(s)

    Ronja Foraita, Fabian Sobotka
    Bremen Institute for Prevention Research and Social Medicine
    (BIPS) http://www.bips.uni-bremen.de

    References

    > MIM (http://www.hypergraph.dk/)
    Edwards D (2002) An Introduction to Graphical Modelling. Springer

    > mimR (http://genetics.agrsci.dk/~sorenh/mimR/index.html)
    Højsgaard S (2004) The mimR package for graphical modelling in R. Journal of Statistical Software, 11(6).

    > CoCo (http://www.badsberg.eu)
    Badsberg JH (2001) A guide to CoCo. Journal of Statistical Software, 6(4).

    > CSI
    Foraita R (2008) A conditional synergy index to assess biological interaction. http://nbn-resolving.de/urn:nbn:de:gbv:46-diss000111139

    See Also

    mimR, CoCo

    Examples

        ### Generates a data frame given a dependence model
        gm.a <- gm.modelsim(1000,"ABC,CDE")
        
        ### Modelselection with graphical output
        gm.analysis(gm.a)   
        
        ### Model validation using the bootstrap 
        gm.boot.coco(100,gm.a,recursive=TRUE,follow=TRUE)
    
        ### Model prediction using cross validation
        ### Example works!
        ## Not run: gm.cv(3,data=gm.a,strategy="f",options="b")
        
        ### Testing interaction on the penetrance scale
        ### using the conditional synergy index (CSI)
        gm.csi(1,2,3,data=gm.a)
    
        ### Testing interaction on a additivity scale
        ### using the synergy index (S)
        gm.si(1,2,3,data=gm.a)   
    
        ### Gamma Coefficient B indpendent D given C
        gm.gamma(2,4,data=gm.a,conditions=3)
    
    

    [Package gmvalid version 1.2 Index]