evaluate {netbenchmark} | R Documentation |
evaluate
compares the inferred network to the true
underlying network for several threshold values and appends the
resulting confusion matrices to the returned object.
evaluate(inf.net,true.net,sym=TRUE,extend=0)
inf.net |
An adjacency matrix representing the inferred network. |
true.net |
An adjacency matrix representing the true underlying network. |
sym |
Logical, make a symmetric evaluation (default = TRUE). |
extend |
Integer, specifying the desired number of links to extend in the network (default=0) |
The first edgelist network inet
is compared to the true
underlying network, tnet
, in order to compute the
metrics of the performance.
If extend is specified, extend
links that network
inet
has set to 0 are added to the inferred network
randomly at the end of the edgelist.
evaluate
returns a matrix with four columns representing
TP,FP,TN,FN
.
These values are computed for each of the predicted links that
should be sorted. Thus, each row of the returned object contains
the confusion matrix as a function of the cutoff in the edgelist.
# Inference inf.net <- cor(syntren300.data) #Evaluate table <- evaluate(inf.net, syntren300.net) table.nosym <- evaluate(inf.net, syntren300.net,sym=FALSE)