Package whistleClassifier.training
Class SequentialTrainingSelector
java.lang.Object
whistleClassifier.training.TrainingSelector
whistleClassifier.training.SequentialTrainingSelector
- Direct Known Subclasses:
PairedSequentialTrainingSelector
Selects groups of fragments into training / test groups
simply by concatenating all the files together from some random start point.
- Author:
- Doug Gillespie
-
Constructor Summary
Constructors -
Method Summary
Modifier and TypeMethodDescriptionint[]
createSections
(int iSpecies, double trainingFraction, int sectionLength, double minFreq, double maxFreq, int minContourLength) protected int[]
createTrainingSection
(TrainingDataGroup trainingDataGroup, double trainingFraction, int sectionLength, double minFreq, double maxFreq, int minContourLength) Create a training and a test data selection, randomly selected from within the data for a single species.int
Methods inherited from class whistleClassifier.training.TrainingSelector
getNumSpecies, getParameterArray, getTotalFragments, getTrainingDataCollection, getWhistleFragmenter, setTrainingDataCollection, setWhistleFragmenter
-
Constructor Details
-
Method Details
-
createSections
public int[] createSections(int iSpecies, double trainingFraction, int sectionLength, double minFreq, double maxFreq, int minContourLength) -
createTrainingSection
protected int[] createTrainingSection(TrainingDataGroup trainingDataGroup, double trainingFraction, int sectionLength, double minFreq, double maxFreq, int minContourLength) Create a training and a test data selection, randomly selected from within the data for a single species.- Parameters:
trainingDataGroup
- data group (one per species)trainingFraction
- fraction of the data to use.sectionLength
- number of fragments in a classification group- Returns:
- 2 element int array of the number of fragments in the training and in the test sections.
-
getNumTestSets
public int getNumTestSets()- Specified by:
getNumTestSets
in classTrainingSelector
-