mr.net {minet} | R Documentation |
mr.net
infers a network using the maximum relevance/minimum redundancy
feature selection method - see details.
mr.net( mim )
mim |
the mutual information matrix - see build.mim . |
Consider a supervised learning task, where the output is denoted by Y
and V is the set of input variables. The method ranks the set
V of inputs according to a score that is the difference between
the mutual information with the output variable Y (maximum relevance)
and the average mutual information with the previously ranked variables
(minimum redundancy).
The greedy search starts by selecting the variable Xi having the highest
mutual information with the target Y. The second selected variable Xj
will be the one that maximizes I(Xj;Y)-I(Xj;Xi).
In the following steps, given a set S of selected variables, the criterion
updates S by choosing the variable Xk that maximizes
I(Xk;Y) - mean(I(Xk;Xi)), Xi in S.
The MRNET approach consists in repeating this selection procedure for
each target variable by putting Y=Xi and
V = X\{Xi},
i=1,...,n where X is the set of outcomes of all variables.
The weight of each pair Xi,Xj will be the maximum score between the one
computed when Xi is the output and the one computed when Xj is
the output.
mr.net
returns a matrix which is the weighted adjacency matrix of the network.
In order to display the network, load the package Rgraphviz and use the following command:
plot( as( returned.matrix ,"graphNEL") )
P.E.Meyer, F.Lafitte, G.Bontempi
Patrick E. Meyer, Kevin Kontos, Frederic Lafitte, and Gianluca Bontempi. Information-theoretic inference of large transcriptional regulatory networks. EURASIP Journal on Bioinformatics and Systems Biology, 2007.
build.mim
, clr.net
, aracne.net
data(syn.data) mim <- build.mim(disc(syn.data)) net <- mr.net(mim)