Package ristretto.tools

Class FSErrorRateValidator

  • java.lang.Object
    • ristretto.tools.FSErrorRateValidator
  • public class FSErrorRateValidator
extends java.lang.Object
    Obtain the training and test error rates of the results of a Feature Selector

    • Method Summary

      All Methods Static Methods Concrete Methods 
      Modifier and Type Method Description
      static void main​(java.lang.String[] args)
      Obtain the training and test error rates of the results of a Feature Selector

      • Methods inherited from class java.lang.Object

        clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    • Constructor Detail

      • FSErrorRateValidator

        public FSErrorRateValidator()

    • Method Detail

      • main

        public static void main​(java.lang.String[] args)
        Obtain the training and test error rates of the results of a Feature Selector

        Expected args:

        1. Path to the training dataset
        2. Path to the test dataset
        3. Class column index (-1 for unlabeled data)

        Optional args:

        • applylda: If this flag appears, LDA is applied to the selected features of the dataset before classifying
        • classifier: Possible values are knn, nbc and svm. If omitted, nbc is used by default
        • selected features: A list of feature indices, if provided, only these features are taken into account. Otherwise all features are used

        Notes:

        • For knn: k is set as sqrt(num_samples) if knn is selected
        • For svm: a RBF kernel is assumed and C and gamma are read from a file named "svm_params.data"
        Parameters:
        args - Command line arguments