rkrnsRanker: use random forest or svm to predict and rank sentences from...
In stnava/RKRNS: ANTsR Decoding

Description Usage Examples

takes a dataframe that contains a column named lab such that mydf$lab gives the true labeling of the data. trainingRows is a vector identifying which rows should be training data. success is defined by a prediction probability being within the top targetSuccessRate*100 percent. Returns various success metrics.

1 2	rkrnsRanker(mydfin, trainingRows, targetSuccessRate = 0.15, whichmodel = "randomForest", verbose = FALSE)

mylabs<-rep(c("a","b","c"),5)
voxeldata<-replicate(100, rnorm(length(mylabs)))
mydf<-data.frame( lab=mylabs, vox=voxeldata)
rfr<-rkrnsRanker( mydf, 1:round(nrow(mydf)/2), whichmodel='svm' )
print( rfr$successPercent )