|Home >> All|
|||marf.Classification.* (19)||||marf.FeatureExtraction.* (10)||||marf.Preprocessing.* (14)|
|||marf.Stats.* (18)||||marf.Storage.* (27)||||marf.gui.* (7)|
|||marf.math.* (4)||||marf.nlp.* (42)||||marf.util.* (16)|
marf: Javadoc index of package marf.
Arrays: marf.util.Arrays is an extension of java.util.Arrays to group a lot of commonly used arrays-related functionality in one place. This class can do whatever java.util.Arrays can, plus allows copying array portions, including cases when the source and destination arrays are of different types, and providing array-to-Vector conversions. For the type-conversion routines a proper casting to the destination type is performed when needed. It also allows inheritance from the class, so that anyone wishing to extend it is welcome to do so without the pain of re-wrapping the methods. NOTE: the java.util.Arrays ...
FreeVector: Adaptive extension of the java.util.Vector class. You may access elements of a Vector beyond it's initial length --- the Vector will be automaticall adjusted as appropriate. Useful in the applications where desirable vector's growth by setting an element beyond its upper boundary automatticaly lengthens the vector to accomondate the change (similar to Perl arrays). Similarly, getting an element beyond the upper boundary is not desirable failure, but an empty element returned. This makes the application to see as vector of a theoretically infinite in length. TODO: allow negative index boundaries. ...
BaseThread: Class BaseThread is customized base class for many of our own threads. It provides an attempt to maintain an automatic unique TID (thread ID) among all the derivatives and allow setting your own if needed, integrates with ExpandedThreadGroup, and maintains a local reference for the Runnable target if clients need it. Java 1.5 NOTE: In this Java version they finally managed to provide a method similar to our getTID() , called getId() and this class was created prior that. And the functionality we offer seems to be superior anyway. $Id: BaseThread.java,v 1.16 2005/06/20 05:36:43 mokhov Exp $
MinMaxAmplitudes: Min/Max Amplitudes. Extracts N minimum and X maximum amplitudes from a sample as features. If incoming sample array's length is less than N + X, it is adjusted to be N + X long with the length/2 value repeated N + X - length times. $Id: MinMaxAmplitudes.java,v 1.12 2005/08/13 23:09:37 susan_fan Exp $ TODO: needs improvement to select different amplitudes as we don't want 20 the same maximums or minimus if others are avaible.
Matrix: Math Matrix Operations. NOTE: this class provides a lot of useful and working functionality, but requires a lot of improvements. In particular in performance, documentation, and styles and consistency in operators. Some missing features will be added/filled in as well. Requires a lot of thorough testing and MathTestApp serves that purpose. $Id: Matrix.java,v 1.30 2005/08/13 23:09:39 susan_fan Exp $
RandomClassification: Random Classification Module is for testing purposes. This represents the bottomline of the classification results. All the other modules should be better than this 99% of the time. If they are not, debug them. $Id: RandomClassification.java,v 1.15 2005/08/14 01:15:55 mokhov Exp $
ExpandedThreadGroup: Provides some useful extensions to java.lang.ThreadGroup one would normally expect to be in a "group". Maintains local references to the group-belonging threads for extra group control in a form of a Vector. $Id: ExpandedThreadGroup.java,v 1.10 2005/06/20 05:34:32 mokhov Exp $
Classification: Abstract Classification Module. A generic implementation of the IClassification interface. The derivatives must inherit from this module, and if they cannot, they should implement IClassification themselves. $Id: Classification.java,v 1.39 2005/08/05 22:19:47 mokhov Exp $
StatsCollector: TODO: Implement. Ideally we'd want to measure. How long it takes for a particular module to do the thing. How long it takes to run whole pipeline. How many features... Amount of noise and silence removed... ... $Id: StatsCollector.java,v 1.7 2005/07/30 20:05:45 mokhov Exp $
MahalanobisDistance: Mahalanobis Distance Classification Module. NOTE : Implemented as equivalent to Euclidean Distance in 0.2.0, i.e. the Covariance matrix is always an Indentity one. $Id: MahalanobisDistance.java,v 1.21 2005/08/12 20:02:56 susan_fan Exp $
Compiler: Compiler Class. (C) 2001-2005 Serguei Mokhov, This class has the main function of the lexical analyzer program which is simply a stub invoking the methods of the lexer. $Id: Compiler.java,v 1.8 2005/06/16 19:58:57 mokhov Exp $
ISampleLoader: Samle loading interface. Must be overriden by a concrete sample loader. Derivatives should try their best to inherit from SampleLoader class; otherwise they must implement this. $Id: ISampleLoader.java,v 1.9 2005/06/12 23:09:37 mokhov Exp $
OptionProcessor: Command-Line Option Processing Facilitating Utility. Helps to maintain and validate command-line options and their arguments. The class is properly synchronized as of 0.3.0.4. $Id: OptionProcessor.java,v 1.29 2005/08/11 00:44:50 mokhov Exp $
ModuleParams: Class ModuleParams provides ability to pass module-specific parameters from an application. The specific module should know in which order and how to downcast those params. $Id: ModuleParams.java,v 1.13 2005/06/16 19:58:54 mokhov Exp $
FeatureExtraction: Generic Feature Extraction Module. Every feature extraction module must extend this class; if it cannot then they must implement the IFeatureExtraction interface. $Id: FeatureExtraction.java,v 1.30 2005/08/05 22:19:54 mokhov Exp $
StorageManager: Class StorageManager. Almost every concrete module must inherit from this class. If that's not possible, implement IStorageManager interface. $Id: StorageManager.java,v 1.23 2005/06/20 21:45:15 mokhov Exp $
Result: Represents a single classification result - ID and some value indicating either certain distance from the sample being recognized or a probability. $Id: Result.java,v 1.20 2005/08/13 16:28:49 mokhov Exp $
FFTFilter: FFTFilter class implements filtering using the FFT algorithm. Derivatives must set frequency response based on the type of filter they are. $Id: FFTFilter.java,v 1.23 2005/08/05 22:19:55 mokhov Exp $
Algorithms: Collection of algorithms to be used by the modules. Decouples algorithms from the modules and allows to be used by different types of modules. $Id: Algorithms.java,v 1.8 2005/08/01 16:27:19 mokhov Exp $
CompilerError: Generic Compiler Error. Normally subclassed to differentiate between various error types like lexical, syntactical, semantic and such. $Id: CompilerError.java,v 1.12 2005/08/13 23:09:39 susan_fan Exp $
TrainingSample: TrainingSample contains one item in the training set. Each training sample consists of the feature vector plus information describing that feature vector. Has been extracted from TrainingSet in 0.3.0.
MaxProbabilityClassifier: Maximum Probability Classification Module. Originally came with the LangIdentApp NLP application of Serguei Mokhov. $Id: MaxProbabilityClassifier.java,v 1.19 2005/08/13 23:09:37 susan_fan Exp $
IStatisticalEstimator: Implements generic Statistical Estimator routines. Must be subclasses by concrete implemenations of statistical estimators. $Id: IStatisticalEstimator.java,v 1.1 2005/06/12 23:09:37 mokhov Exp $
StatisticalEstimator: Implements generic Statistical Estimator routines. Must be subclasses by concrete implemenations of statistical estimators. $Id: StatisticalEstimator.java,v 1.24 2005/06/16 19:58:47 mokhov Exp $
Stemming: General Stemmer. Must be subclassed by a language-specific stemmers. If they can't, they must implement IStemming. TODO: implement. $Id: Stemming.java,v 1.11 2005/06/16 19:58:57 mokhov Exp $