All Downloads are FREE. Search and download functionalities are using the official Maven repository.

edu.ucla.sspace.mains.LSAMain Maven / Gradle / Ivy

Go to download

The S-Space Package is a collection of algorithms for building Semantic Spaces as well as a highly-scalable library for designing new distributional semantics algorithms. Distributional algorithms process text corpora and represent the semantic for words as high dimensional feature vectors. This package also includes matrices, vectors, and numerous clustering algorithms. These approaches are known by many names, such as word spaces, semantic spaces, or distributed semantics and rest upon the Distributional Hypothesis: words that appear in similar contexts have similar meanings.

The newest version!
/*
 * Copyright 2009 David Jurgens
 *
 * This file is part of the S-Space package and is covered under the terms and
 * conditions therein.
 *
 * The S-Space package is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as published
 * by the Free Software Foundation and distributed hereunder to you.
 *
 * THIS SOFTWARE IS PROVIDED "AS IS" AND NO REPRESENTATIONS OR WARRANTIES,
 * EXPRESS OR IMPLIED ARE MADE.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, WE MAKE
 * NO REPRESENTATIONS OR WARRANTIES OF MERCHANT- ABILITY OR FITNESS FOR ANY
 * PARTICULAR PURPOSE OR THAT THE USE OF THE LICENSED SOFTWARE OR DOCUMENTATION
 * WILL NOT INFRINGE ANY THIRD PARTY PATENTS, COPYRIGHTS, TRADEMARKS OR OTHER
 * RIGHTS.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program. If not, see .
 */

package edu.ucla.sspace.mains;

import edu.ucla.sspace.basis.BasisMapping;
import edu.ucla.sspace.basis.StringBasisMapping;

import edu.ucla.sspace.common.ArgOptions;
import edu.ucla.sspace.common.SemanticSpace;
import edu.ucla.sspace.common.SemanticSpaceIO.SSpaceFormat;

import edu.ucla.sspace.lsa.LatentSemanticAnalysis;

import edu.ucla.sspace.matrix.LogEntropyTransform;
import edu.ucla.sspace.matrix.MatrixFactorization;
import edu.ucla.sspace.matrix.Transform;
import edu.ucla.sspace.matrix.SVD;
import edu.ucla.sspace.matrix.SVD.Algorithm;

import edu.ucla.sspace.util.ReflectionUtil;
import edu.ucla.sspace.util.SerializableUtil;

import java.io.IOError;
import java.io.IOException;

import java.util.concurrent.ConcurrentHashMap;


/**
 * An executable class for running {@link LatentSemanticAnalysis} (LSA) from the
 * command line.  This class takes in several command line arguments.
 *
 * 
    * *
  • Required (at least one of): *
      * *
    • {@code -d}, {@code --docFile=FILE[,FILE...]} a file where each line is * a document. This is the preferred input format for large corpora * *
    • {@code -f}, {@code --fileList=FILE[,FILE...]} a list of document files * where each file is specified on its own line. * *
    * *
  • Algorithm Options: *
      * *
    • {@code --dimensions=} how many dimensions to use for the LSA * vectors. See {@link LatentSemanticAnalysis} for default value * *
    • {@code --preprocess=} specifies an instance of {@link * edu.ucla.sspace.lsa.MatrixTransformer} to use in preprocessing the * word-document matrix compiled by LSA prior to computing the SVD. See * {@link LatentSemanticAnalysis} for default value * *
    • {@code -F}, {@code --tokenFilter=FILE[include|exclude][,FILE...]} * specifies a list of one or more files to use for {@link * edu.ucla.sspace.text.TokenFilter filtering} the documents. An option * flag may be added to each file to specify how the words in the filter * filter should be used: {@code include} if only the words in the filter * file should be retained in the document; {@code exclude} if only the * words not in the filter file should be retained in the * document. * *
    • {@code -S}, {@code --svdAlgorithm}={@link * edu.ucla.sspace.matrix.SVD.Algorithm} species a specific {@code * SVD.Algorithm} method to use when reducing the dimensionality in LSA. * In general, users should not need to specify this option, as the * default setting will choose the fastest algorithm available on the * system. This is only provided as an advanced option for users who * want to compare the algorithms' performance or any variations between * the SVD results. * *
    * *
  • Program Options: *
      * *
    • {@code -o}, {@code --outputFormat=}text|binary} Specifies the * output formatting to use when generating the semantic space ({@code * .sspace}) file. See {@link edu.ucla.sspace.common.SemanticSpaceUtils * SemanticSpaceUtils} for format details. * *
    • {@code -t}, {@code --threads=INT} how many threads to use when * processing the documents. The default is one per core. * *
    • {@code -w}, {@code --overwrite=BOOL} specifies whether to overwrite * the existing output files. The default is {@code true}. If set to * {@code false}, a unique integer is inserted into the file name. * *
    • {@code -v}, {@code --verbose} specifies whether to print runtime * information to standard out * *
    * *
* *

* * An invocation will produce one file as output {@code * lsa-semantic-space.sspace}. If {@code overwrite} was set to {@code true}, * this file will be replaced for each new semantic space. Otherwise, a new * output file of the format {@code lsa-semantic-space.sspace} will be * created, where {@code } is a unique identifier for that program's * invocation. The output file will be placed in the directory specified on the * command line. * *

* * This class is desgined to run multi-threaded and performs well with one * thread per core, which is the default setting. * * @see LatentSemanticAnalysis * @see edu.ucla.sspace.matrix.Transform Transform * * @author David Jurgens */ public class LSAMain extends GenericMain { private BasisMapping basis; private LSAMain() { } /** * Adds all of the options to the {@link ArgOptions}. */ protected void addExtraOptions(ArgOptions options) { options.addOption('n', "dimensions", "the number of dimensions in the semantic space", true, "INT", "Algorithm Options"); options.addOption('p', "preprocess", "a MatrixTransform class to " + "use for preprocessing", true, "CLASSNAME", "Algorithm Options"); options.addOption('S', "svdAlgorithm", "a specific SVD algorithm to use" , true, "SVD.Algorithm", "Advanced Algorithm Options"); options.addOption('B', "saveTermBasis", "If true, the term basis mapping will be stored " + "to the given file name", true, "FILE", "Optional"); } public static void main(String[] args) throws Exception { LSAMain lsa = new LSAMain(); lsa.run(args); } protected SemanticSpace getSpace() { try { int dimensions = argOptions.getIntOption("dimensions", 300); Transform transform = new LogEntropyTransform(); if (argOptions.hasOption("preprocess")) transform = ReflectionUtil.getObjectInstance( argOptions.getStringOption("preprocess")); String algName = argOptions.getStringOption("svdAlgorithm", "ANY"); MatrixFactorization factorization = SVD.getFactorization( Algorithm.valueOf(algName.toUpperCase())); basis = new StringBasisMapping(); return new LatentSemanticAnalysis( false, dimensions, transform, factorization, false, basis); } catch (IOException ioe) { throw new IOError(ioe); } } /** * Returns the {@likn SSpaceFormat.BINARY binary} format as the default * format of a {@code LatentSemanticAnalysis} space. */ protected SSpaceFormat getSpaceFormat() { return SSpaceFormat.BINARY; } protected void postProcessing() { if (argOptions.hasOption('B')) SerializableUtil.save(basis, argOptions.getStringOption('B')); } /** * {@inheritDoc} */ protected String getAlgorithmSpecifics() { return "The --svdAlgorithm provides a way to manually specify which " + "algorithm should\nbe used internally. This option should not be" + " used normally, as LSA will\nselect the fastest algorithm " + "available. However, in the event that it\nis needed, valid" + " options are: SVDLIBC, SVDLIBJ, MATLAB, OCTAVE, JAMA and COLT\n"; } }





© 2015 - 2024 Weber Informatics LLC | Privacy Policy