BDgraph (version 2.3)

compare: Comparing the result according to the true graph

Description

According to the true graph structure, with "compare" function we can compare the result for bdmcmc algorithm.

Usage

compare(est.g, true.g)

Arguments

est.g

adjacency matrix for estimated graph in which $a_{ij}=1$ if there is a link between notes $i$ and $j$, otherwise $a_{ij}=0$.

true.g

adjacency for the true graph in which $a_{ij}=1$ if there is a link between notes $i$ and $j$, otherwise $a_{ij}=0$.

Value

• Returns an output of the BDMCMC algorithm which is list-like and contains the following:
• true positive
  the number of correctly estimated links.
• true negative
  the number of true non-existing links which is correctly estimated.
• false positive
  the number of links which they are not in the true graph, but are incorrectly estimated.
• false negative
  the number of links which they are in the true graph, but are not estimated.
• accuracy
  the number of true results (both true positives and true negatives) divided by the total number of true and false results.
• balanced F-score
  a weighted average of the "positive predictive value" and "true positive rate". F-score value reaches its best value at 1 and worst score at 0.
• positive predictive value
  the number of correctly estimated links divided by the total number of links in the estimated graph.
• true positive rate
  the number of correctly estimated links divided by the total number of links in the true graph.
• false positive rate
  the false positive value divided by the total number of links in the true graph.

References

Mohammadi, A. and E. C. Wit (2012). Gaussian graphical model determination based on birth-death MCMC inference, arXiv:1210.5371v4. http://arxiv.org/abs/1210.5371v4

See Also

bdmcmc and select.g

Examples

p <- 8 # number of nodes 
  # "truK" is the precision matrix of true graph
  truK <- diag(p)
  for (i in 1:(p-1)) truK[i,i+1] <- truK[i+1,i] <- 0.5
  truK[1,p] <- truK[p,1] <- 0.4
  truK # precision matrix of the true graph
  
  # generate the data (200 observations) from multivariate normal 
  # distribution with mean zero and percision matrix "truK"
  data <- mvrnorm(200, c(rep(0,p)), solve(truK))  
  
  # selecting the best graph according to bdmcmc algorithm
  output <- bdmcmc(data, meanzero = T, iter = 5000)
  bdmcmc.g <- select.g(output)

  # selecting the best graph according to glasso by using huge package
  huge.g <- huge(data, method = "glasso")
  huge.g <- huge.select(huge.g, criterion = "stars")    
  
  true.g <- ceiling (truK)
  true.g # matrix A shows circle graph with 8 links and 8 nodes which is the true graphical model
  
  # comparing the result by using "compare" function
  compare.true <- compare(est.g = true.g, true.g = true.g)
  compare.bdmcmc <- compare(est.g = bdmcmc.g, true.g = true.g)
  compare.huge <- compare(est.g = huge.g$refit, true.g = true.g)  
  compare.all <- cbind(compare.true, compare.bdmcmc, compare.huge)
  colnames(compare.all) <- c("True graph", "BDgraph", "huge")
  compare.all