Documentation of 'org.encog.util.normalize.segregate.index.IndexSampleSegregator' Java class
IndexSampleSegregator
org.encog.util.normalize.segregate.index

Class IndexSampleSegregator

  • All Implemented Interfaces:
    java.io.Serializable, Segregator


    public class IndexSampleSegregator
    extends IndexSegregator
    An index segregator is used to segregate the data according to its index. Nothing about the data is actually compared. This makes the index range segregator very useful for breaking the data into training and validation sets. For example, you could very easily determine that 70% of the data is for training, and 30% for validation. This segregator takes a starting and ending index, as well as a smple size. Everything that is between these two indexes will be used. The sample repeats over and over. For example, if you choose a sample size of 10, and a beginning index of 0 and an ending index of 5, you would get half of the first 10 element, then half of the next ten, and so on.
    See Also:
    Serialized Form
    • Constructor Detail

      • IndexSampleSegregator

        public IndexSampleSegregator()
        The default constructor, for reflection.
      • IndexSampleSegregator

        public IndexSampleSegregator(int startingIndex,
                                     int endingIndex,
                                     int sampleSize)
        Construct an index sample segregator.
        Parameters:
        startingIndex - The starting index.
        endingIndex - The ending index.
        sampleSize - The sample size.
    • Method Detail

      • getEndingIndex

        public int getEndingIndex()
        Returns:
        The ending index.
      • getSampleSize

        public int getSampleSize()
        Returns:
        The sample size.
      • getStartingIndex

        public int getStartingIndex()
        Returns:
        The starting index.
      • shouldInclude

        public boolean shouldInclude()
        Should this row be included.
        Returns:
        True if this row should be included.

DMelt 3.0 © DataMelt by jWork.ORG

You see the box below because you did not login.