BANFF (version 1.1)

SummaryAccuracy: Summary of the Accuracy of Node and Network selection

Description

This function is to summarize the accuracy of node and Network selection achieved by the functions Networks.STD and Networks.Fast. This function is designed to validate whether the settings of arguments are appropriate for one simulation study and provide insights of setting arguments for real data.

Usage

SummaryAccuracy(Trace,No.Sets,
Type.Accuracy=c("Node","Sub-network"),True.Node,
TruePositive.Net,FalsePositive.Net,
Type.Net.Accuracy=c("Marginal","Sample"),Tolerance)

Arguments

Trace

the trace (results of each iteration) generated by the functions Networks.STD() and Networks.Fast(). For Networks.STD() and Networks.Fast(), set Trace=total$trace where total is the output of the two functions, Networks.STD() and Networks.Fast(). Use rbind() bind the inputs if you have multiple sets of results

No.Sets

number of sets of results you put in the argument Trace

Type.Accuracy

character taking value "Node" or "Sub-network". "Node" represents checking the accuracy of each node; "Sub-network" represents checking the accuracy of network of interest

True.Node

a binary(0/1) vector providing the information of each node. The length of vector should be equal to the length of the vector of pvalue

TruePositive.Net

a vector representing the network of interest consisted of assumed positive nodes. For example, if you assume No.6,No.8 and No.9 nodes consist the network of interest, set TruePositive.Net=c(6,8,9)

FalsePositive.Net

a vector representing the assumed negative nodes which would lead false positive networking (larger network) if such nodes are assumed positive in the results. Such nodes should be connected to the network of interest and assumed negative. For example, if you assume No.3 and No.11 nodes would lead false positive networking (larger network), set TruePositive.Net=c(3,11)

Type.Net.Accuracy

character taking value "Marginal" or "Sample". "Marginal" represents using marginal distribution of each data-set for checking accuracy, for example, if the probability (mean) of one node of a data-set is larger than 0.5, we consider it positive, otherwise, negative. Setting "Sample" represents checking accuracy through each iteration. For Type.Accuracy="Node", only when Type.Net.Accuracy="Sample", is valid.

Tolerance

percentage of the Tolerance rate of accuracy. This argument is only valid when set Type.Accuracy="Networking". Setting Tolerance=0.9 means we consider selecting a correct networking selection if more than 0.9 positive nodes of networking or TruePositive.Net are considered as positive AND less than 0.1 negative nodes of networking or FalsePositive.Net are considered as positive.

Value

TPR.average

The Average of True Positive Rate

FPR.average

The Average of True False Rate

FDR.average

The Average of False Positive Rate

TPR

True Positive Rate of each node

FPR

False Positive Rate of each node

References

Yize Zhao, Jian Kang, Tianwei Yu (2014) A Bayesian nonparameteric model for selecting gene and gene sub network, Annals of Applied Statistics, in press.

Zhou Lan, Jian Kang, Tianwei Yu, Yize Zhao, BANFF: an R package for network identifications via Bayesian nonparametric mixture models, working paper.

Examples

####Gene Network discovery
##Generating Scale free Gene Network
library(igraph)
library(BANFF)
g <- barabasi.game(50, power=1, zero.appeal=1.5,directed = FALSE)
net=as(get.adjacency(g,attr=NULL),"matrix")
##Random assign selected genes and make the signal intension as gaussian mixture
newz=rep(c(1,0,0,1,0),10)
Simnorm=function(n){
weight = c(0.4, 0.6)
mu = c(8,6)
sigma = c(1,0.5)
z = sample(c(1,2),size=n, prob=weight,replace=TRUE)
r = rnorm(n,mean=mu[z],sd=sigma[z])
return(r)
}
testcov<-0
for(i in 1:50){
 if(newz[i]==0){
   testcov[i]<-rnorm(1,mean=0,sd=1)
 }else{
  testcov[i]<-Simnorm(1)
 }
}
pvalue=pnorm(-testcov)
total1=Networks.Fast(pvalue,net,iter=5,nburns=2,
v=20,algorithms="DPM",
piall=c(0.8, 0.85, 0.9, 0.95),rhoall=c(1, 2, 5, 10, 15))
####checking accuracy
###checking accuracy of nodes
SummaryAccuracy(Trace=total1$trace,No.Sets=1,Type.Accuracy="Node",
True.Node=newz,Type.Net.Accuracy="Sample")