Package edu.stanford.nlp.maxent

This package deals with defining and solving maximum entropy problems.

See:
          Description

Class Summary
BinaryFeature This is used when only binary features are needed.
CGRunner This class will call Conjugate Gradient on a LambdaSolve object to find optimal parameters, including imposing a Gaussian prior on those parameters.
Convert This is used to convert an array of double into byte array which makes it possible to keep it more efficiently.
DataGeneric A class representing a data item with an array of inputs X and a String classification Y
Experiments This class represents the training samples.
Feature This class is used as a base class for TaggerFeature for the tagging problem and for BinaryFeature for the general problem with binary features.
Features An ArrayList of Feature
Problem This is a general class for Problem to be solved by the MaxEnt toolkit.
 

Package edu.stanford.nlp.maxent Description

This package deals with defining and solving maximum entropy problems. In the future it will have facilities for easier definition of maxent problems.

If you are new to this package, take a look at the following classes:

Possibly the simplest way to use it is to fill up a Type2Dataset with Type2Datum objects (a Type2Datum is essentially a map from classes into sets of feature values), and then to use LinearType2Classifier.trainClassifier() on your Type2Dataset to train a classifier. You can then use the classOf(), scoresOf(), and justificationOf() methods of the resulting LinearType2Classifier object.