Top Description Interfaces Classes
[default module]

Package javax.visrec.ml.classification

Classification represents task of determining a category of a given object. Machine learning classifiers are able to learn from examples how to determine a category (or a label) for a given input object, that has not seen before (although it should be similar to examples that were used to build a classifier).

Interface Summary

Modifier and TypeInterface and Description
public interface
BinaryClassifier<T>

Binary classifier classifies input object into one of two possible categories (for example: true/false, yes/no, red/blue, positive/negative, spam/not-spam, fraud/not-fraud).

public interface
Classifiable<
Type of input for classifier
T,
Type of categories/labels (could be anything like enum, String, Integer or user defined class)
C>

Classes that implement this interface enable direct classification of Java objects, no need to manually convert to data structures used internally by ML.

public interface
Classifier<
type of input objects to classify (eg. User, Product, Transaction, Image, etc.)
T,
type of classification result (String, Enum, custom class).
R>

Generic classifier interface, that all classifiers should implement, and it provides a method to classify given instances of some class.

public interface
ImageClassifier<
type of input objects to classify (eg. User, Product, Transaction, Image, etc.)
IMAGE_CLASS>

Classifier interface specialized in image classification

public interface
MultiClassClassifier<
Type of input objects (which are being classified)
T,
Type of classifier return value - type of object which represent category class.
R>

Machine learning algorithms that provide multi class classification.

public interface
NeuralNetBinaryClassifier<T>

public interface
NeuralNetImageClassifier<T>

Class Summary

Modifier and TypeClass and Description
public abstract class
AbstractMultiClassClassifier<
class of machine learning model back-end
MODEL_CLASS,
type of classifier's input
T,
type of classifier's results/class
R>

Skeleton base class for implementations of multi class classifiers.

public class
ClassificationException

Exception thrown if anything fails in the execution of a classifier.

public class
EnsambleClassifier<
The input type which is to be classified.
T,
Return type of the classifier.
R>

EnsambleClassifier is a group of classifiers that provide common classification result, which gives better accuracy then each individual classifier.

public abstract class
LogisticRegression<
Implementation class of underlying machine learning model
MODEL_CLASS>

This class performs basic binary classification - mapping of specified input to true/false with probability using logistic regression algorithm.