Skip to main content

Infer.NET user guide : Learners : Bayes Point Machine classifiers : Command-line runners


You can use the CrossValidate module to assess the generalization performance of the Bayes Point Machine, both in binary and multi-class classification. The CrossValidate module starts by reading a labelled data from a file and partitions its instances into K subsets of equal size, known as folds. It then trains the Bayes Point Machine classifier on K - 1 folds and evaluates its performance on the withheld K-th fold. It cycles through all K combinations of splits into training and validation sets to finally report the overall performance results.

The CrossValidate module has the following command-line arguments:

Required arguments

Optional arguments

For more information about the command-line arguments, see Settings.


Learner Classifier BinaryBayesPointMachine CrossValidate   
    --data-set training.dat --results cross-validation-results.csv   
    --iterations 15 --batches 1 --compute-evidence  

Learner Classifier MulticlassBayesPointMachine CrossValidate   
    --data-set training.dat --results cross-validation-results.csv   
    --iterations 15 --batches 1 --compute-evidence