net.perf: Link Prediction Performance

Description Usage Arguments Details Value Examples

View source: R/netpredictor.R

Description

This function samples links and removies links from the adjacency matrix and predicts them and calculates area under accumulation curve, AUC, bedroc, and Enrichment factor.

Usage

1
2
3
net.perf(A, S1, S2, restart = 0.8, alpha = 0.5, lamda = 0.5,
  relinks = 100, numT = 2, norm = "laplace", Calgo = c("rwr", "nbi",
  "netcombo", "all"))

Arguments

A

Drug target association matrix

S1

Sequence similarity matrix object

S2

Accepts a matrix object similarity scores for compounds.

restart

restart value if using rwr or netcombo

alpha

alpha value if using nbi or netcombo

lamda

lamda value if using nbi or netcombo

relinks

Number of links to remove randomly from the input matrix.

numT

Frequency of the number of targets.

norm

normalization of matrices options are "laplace" or "none".

Calgo

Algorithm to use for Bipartite link prediction options are "rwr","nbi" & "netcombo".

Details

get the performance of the link Prediction algorithms.

Value

it returns a list of aucc,auc, bedorc,enrichment factor and auc (top 10

Examples

1
2
3
4
5
6
7
8
## Not run: 
data(Enzyme)
A = enzyme_ADJ
S1 = enzyme_Gsim
S2= enzyme_Csim
m = net.perf(A,S1,S2,alpha=0.5,lamda=0.5,relinks = 50,numT=2,norm="laplace",Calgo="nbi")

## End(Not run)

abhik1368/netpredictor documentation built on May 10, 2017, 9:09 p.m.