confusion_proba_plot: Plot the distribution of signals in categories after...
In jiho/soundclass: Signal classifier for sounds

Description Usage Arguments See Also Examples

Plot the distribution of signals in categories after prediction

1	confusion_proba_plot(true, pred, prob)

`true`	true categories
`pred`	predicted categories (last column of the data.frame output by `predict.gbdt`)
`prob`	for each signal, probabilities to be in each category
`...`	passed to `geom_??`

gbm, summary.gbm, predict.gbm in package gbm

data(sirena)
sub <- subsample(sirena[,-ncol(sirena)], p=0.2)
train <- sirena[sub$picked,]
data <- sirena[!sub$picked,]
p <- classify(data=data[,-ncol(data)], train=train,
              n.trees=100, shrinkage=0.01, n.minobsinnode=1,
              n.cores=1)
confusion_proba_plot(true=data$type, pred=p$type, prob=p[,1:4])