Predictions
We could also get the predictions on the test dataset using the getBMRPredictions function. The two tables in this section show the actual and the predicted labels of a few images represented by the ID column. Observe that the predictions are not perfect, just as we would expect from the relatively low overall accuracy.
Predictions using randomForestSRC:
head(getBMRPredictions(bmr, as.df = TRUE))
Figure 9.13: The actual labels.
Figure 9.14: The predicted labels.
Learners and measures
The getBMRLearners function gives details about the learners used in the benchmark. Information such as hyperparameter and predict-type could be obtained using this function. Similarly, the getBMRMeasures function provides details such as best about the performance measures. The following table shows the details about the measures we used in our benchmark experiment:
getBMRLearners(bmr)
The output is as follows:
## $multilabel.randomForestSRC ## Learner multilabel.randomForestSRC from package randomForestSRC...
