Description Usage Arguments Value Examples
View source: R/VNviaCANsample.R
Semi-supervised Vertex Nomination using Canonical Sampling
1 | VNviaCANsample(n, m, Lam, A, observe, truth, numburn, numsample)
|
n |
Vector of number of unsupervised datapoints per block. |
m |
Vector of number of supervsised datapoints per block. |
Lam |
Interblock link probability matrix |
A |
Adjacency matrix (undirected, hollow) |
observe |
length sum(n+m) vector of known block identities. If -1, unknown. If >=1, that is the block identity. |
truth |
length sum(n+m) vector of true block identities. |
numburn |
Integer number of swaps to perform before sampling occurs. |
numsample |
Integer number of swaps to perform while sampling occurs. |
probs = probs,avgprec = avgprec,reveal = reveal
probs |
Length sum(n) vector of estimated probability of block 1 membership, sorted in nomination order |
avgprec |
Average precision (at n(1)). Higher is better. 1 is the best. |
reveal |
boolean vector of length sum(n) of whether or not the i^th vector in the nomination order is from block 1. |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 | library(ssClust)
numVert=500 #number of ambiguous vertices
#matching Li's simulation parameters
rho <- c(0.4,0.3,.3)
B1 <- matrix(c(0.5,0.3,.4,
0.3,0.8,.6,
.4,0.6, .3), nrow = 3,ncol=3)
B2 <- matrix(.5 ,nrow = 3,ncol=3)
nu = .3
Lam0 = nu*B1 + (1-nu)*B2
n0 <- numVert*rho
m0<-c(20, 0, 0)
numburn0 = 10^4
numsample0 = 10^4
sbm.out = makeSBM(n0,m0,Lam0)
A0 = as.matrix(sbm.out$A)
observe0 = sbm.out$observe
truth0 = sbm.out$truth
canSamp.out = VNviaCANsample(n=n0, m=m0, Lam=Lam0, A=A0,
observe=observe0,
truth=truth0,
numburn=numburn0,
numsample=numsample0)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.