svmtrain: False positive rates from several 1-SVM models
In directlabels: Direct Labels for Multicolor Plots

svmtrain

R Documentation

False positive rates from several 1-SVM models

Support Vector Machine density estimation (1-SVM) was applied to a set of negative control samples, and then used to test on a positive control.

data(svmtrain)

A data frame with 378 observations on the following 5 variables.

replicate: a factor with levels 1 2 3, the experimental replicate. We fit 1-SVM models to each replicate separately.
rate: a numeric vector, the percent of observations that were outside the trained model.
data: a factor with levels KIF11 test train, which set of observations did we measure. test and train are each 50% random splits of the negative controls in the experiment, and KIF11 is the positive control in the experiment.
gamma: a numeric vector, the tuning parameter of the radial basis function kernel.
nu: a numeric vector, the regularization parameter of the 1-SVM.