SVMPredict: Predict false interactions using a training set

In ppiPre: Predict Protein-Protein Interactions Based on Functional and Topological Similarities

Description

Using a training set to train the SVM classifier, and predict false interactions in a user given PPI network.

Usage

SVMPredict(training_set,predict_set,output="falsePPIs-ppiPre.csv",organism="yeast",
           drop ="IEA", replaceNA=0)

Arguments

training_set	CSV format golden standard training set
predict_set	PPI network to do the prediction
output	CSV format file to ave the result
organism	One of "anopheles", "arabidopsis", "bovine", "canine", "chicken", "chimp", "ecolik12", "ecsakai", "fly", "human", "malaria", "mouse", "pig", "rat", "rhesus", "worm", "xenopus", "yeast" and "zebrafish."
drop	A set of evidence codes based on which certain annotations are dropped. Use NULL to keep all GO annotations.
replaceNA	The value to replace NA in training and predict set.

Author(s)

Yue Deng <[e-mail]>

See Also

TopologicSims GOKEGGSims ComputeAllEvidences

Examples

  #edges <- data.frame(node1=c("1132", "1133", "1134", "1134", "1145", "1147"),
  #                    node2=c("1134", "1134", "1145", "1147", "1147", "1149"),
  #                    label=c(1, 1, 1, 0, 0, 0))
  #graph<-graph.data.frame(edges,directed=FALSE)
  #trainsample <- "ppiPre-SVMPredict-trainsample.csv"
  #write.csv(edges,file=trainsample,row.names=FALSE) 
  #edges <- data.frame(node1=c("1132", "1133", "1134", "1134", "1146", "1147"),
  #                    node2=c("1133", "1134", "1142", "1147", "1147", "1149"),
  #                    label=c(1, 0, 1, 0, 1, 0))
  #graph<-igraph::graph.data.frame(edges,directed=FALSE)
  #predictsample <- "ppiPre-SVMPredict-predictsample.csv"
  #write.csv(edges,file=predictsample,row.names=FALSE) 
  #SVMPredict(trainsample, predictsample, organism="human", replaceNA=0)

ppiPre documentation built on May 30, 2017, 4:33 a.m.