Last data update: 2014.03.03

 R: Model selection for nested effect models
nemModelSelectionR Documentation

Model selection for nested effect models

Description

Infers models with different regularization constants, compares them via the BIC or AIC criterion and returns the highest scoring one

Usage

nemModelSelection(lambdas,D,inference="nem.greedy",models=NULL,control=set.default.parameters(unique(colnames(D))),verbose=TRUE,...)

Arguments

lambdas

vector of regularization constants

D

data matrix with experiments in the columns (binary or continious)

inference

search to use exhaustive enumeration, triples for triple-based inference, pairwise for the pairwise heuristic, ModuleNetwork for the module based inference, nem.greedy for greedy hillclimbing, nem.greedyMAP for alternating MAP optimization using log odds or log p-value densities

models

a list of adjacency matrices for model search. If NULL, an exhaustive enumeration of all possible models is performed.

control

list of parameters: see set.default.parameters

verbose

do you want to see progression statements? Default: TRUE

...

other arguments to pass to function nem or network.AIC

Details

nemModelSelection internally calls nem to infer a model with a given regularization constant. The comparison between models is based on the BIC or AIC criterion, depending on the parameters passed to network.AIC.

Value

nem object

Author(s)

Holger Froehlich

See Also

set.default.parameters, nem, network.AIC

Examples

   data("BoutrosRNAi2002")
   D <- BoutrosRNAiDiscrete[,9:16]   
   hyper = set.default.parameters(unique(colnames(D)), para=c(0.13, 0.05), Pm=diag(4))
   res <- nemModelSelection(c(0.1,1,10), D, control=hyper)      
   
   plot.nem(res,main="highest scoring model")

Results


R version 3.3.1 (2016-06-21) -- "Bug in Your Hair"
Copyright (C) 2016 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> library(nem)
> png(filename="/home/ddbj/snapshot/RGM3/R_BC/result/nem/nemModelSelection.Rd_%03d_medium.png", width=480, height=480)
> ### Name: nemModelSelection
> ### Title: Model selection for nested effect models
> ### Aliases: nemModelSelection
> ### Keywords: graphs models
> 
> ### ** Examples
> 
>    data("BoutrosRNAi2002")
>    D <- BoutrosRNAiDiscrete[,9:16]   
>    hyper = set.default.parameters(unique(colnames(D)), para=c(0.13, 0.05), Pm=diag(4))
>    res <- nemModelSelection(c(0.1,1,10), D, control=hyper)      
Greedy hillclimber for 4 S-genes (lambda = 0.1 )...

Computing (marginal) likelihood for 1 models
--> Using regularization to incorporate prior knowledge
12  local models to test ...
Computing (marginal) likelihood for 12 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 20.75064 )
--> Edge added, removed or reversed
12  local models to test ...
Computing (marginal) likelihood for 12 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 1.486884 )
--> Edge added, removed or reversed
11  local models to test ...
Computing (marginal) likelihood for 11 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 0.4669048 )
--> Edge added, removed or reversed
10  local models to test ...
Computing (marginal) likelihood for 10 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 15.02184 )
--> Edge added, removed or reversed
5  local models to test ...
Computing (marginal) likelihood for 5 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 82.00684 )
log-likelihood of model =  -288.0096 
Greedy hillclimber for 4 S-genes (lambda = 1 )...

Computing (marginal) likelihood for 1 models
--> Using regularization to incorporate prior knowledge
12  local models to test ...
Computing (marginal) likelihood for 12 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 20.75064 )
--> Edge added, removed or reversed
12  local models to test ...
Computing (marginal) likelihood for 12 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 1.486884 )
--> Edge added, removed or reversed
11  local models to test ...
Computing (marginal) likelihood for 11 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 1.366905 )
--> Edge added, removed or reversed
10  local models to test ...
Computing (marginal) likelihood for 10 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 13.22184 )
--> Edge added, removed or reversed
5  local models to test ...
Computing (marginal) likelihood for 5 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 83.80684 )
log-likelihood of model =  -255.6682 
Greedy hillclimber for 4 S-genes (lambda = 10 )...

Computing (marginal) likelihood for 1 models
--> Using regularization to incorporate prior knowledge
12  local models to test ...
Computing (marginal) likelihood for 12 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 20.75064 )
--> Edge added, removed or reversed
12  local models to test ...
Computing (marginal) likelihood for 12 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 1.486884 )
--> Edge added, removed or reversed
11  local models to test ...
Computing (marginal) likelihood for 11 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 10.3669 )
--> Edge added, removed or reversed
10  local models to test ...
Computing (marginal) likelihood for 10 models
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
--> Using regularization to incorporate prior knowledge
((Marginal) posterior likelihood difference of best vs. second best model for  rel key tak mkk4hep : 4.778158 )
log-likelihood of model =  -259.0487 
==> AIC ( lambda =   ) =  487.532817873473 ( #param = 5 )===============
==> AIC ( lambda =   ) =  487.532817873473 ( #param = 5 )===============
==> AIC ( lambda =   ) =  509.543078495049 ( #param = 3 )===============
====> chosen best model with lambda = 0.1 
>    
>    plot.nem(res,main="highest scoring model")      
> 
> 
> 
> 
> 
> dev.off()
null device 
          1 
>


Created & Maintained by Osamu Ogasawara (osamu.ogasawara@gmail.com) and IMSBIO