edu.ucla.sspace.wordsi

Class RandomIndexingContextGenerator

java.lang.Object

edu.ucla.sspace.wordsi.RandomIndexingContextGenerator

All Implemented Interfaces:

ContextGenerator

public class RandomIndexingContextGenerator
extends Object
implements ContextGenerator

A random indexing based ContextGenerator. This generator creates a context vector by summing the index vectors associated with co-occurring words. When in normal mode, an index vector is generated for every observed word. When in read only mode, no new index vector are generated. If a PermutationFunction is provided, the index vectors will be permuted based on the distance between the co-occurring word and the focus word in the context.

Author:

Keith Stevens

See Also:

RandomIndexing

Constructor Summary

Constructors

Constructor and Description
RandomIndexingContextGenerator(Map<String,TernaryVector> indexMap, PermutationFunction<TernaryVector> perm, int indexVectorLength) Creates a new RandomIndexingContextGenerator.

Method Summary

Methods

Modifier and Type	Method and Description
protected void	addContextTerms(SparseDoubleVector meaning, Queue<String> words, int distance) Adds the index vector for each co-occurring word in the context.
SparseDoubleVector	generateContext(Queue<String> prevWords, Queue<String> nextWords) Returns a SparseDoubleVector that represents the context composed of the set of prevWords before the focus word and the set of nextWords after the focus word.
int	getVectorLength() Returns the maximum number of dimensions used to represent any given context.
void	setReadOnly(boolean readOnly) Sets the read only mode of the ContextGenerator.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

RandomIndexingContextGenerator

public RandomIndexingContextGenerator(Map<String,TernaryVector> indexMap,
                              PermutationFunction<TernaryVector> perm,
                              int indexVectorLength)

Creates a new RandomIndexingContextGenerator.

Parameters:

indexMap - The map responsible for mapping co-occurring terms to their index vectors. This map should generate index vectors for words that are not currently mapped when in normal mode.

perm - A PermutationFunction for TernaryVectors.

indexVectorLength - The number of dimensions in each index vector.

Method Detail

generateContext

public SparseDoubleVector generateContext(Queue<String> prevWords,
                                 Queue<String> nextWords)

Returns a SparseDoubleVector that represents the context composed of the set of prevWords before the focus word and the set of nextWords after the focus word. Since sparse vectors are returned, if a second order vector is generated, it is recommended that the vector also be sparsed or have very few dimensions.

Specified by:

generateContext in interface ContextGenerator

getVectorLength

public int getVectorLength()

Returns the maximum number of dimensions used to represent any given context.

Specified by:

getVectorLength in interface ContextGenerator

setReadOnly

public void setReadOnly(boolean readOnly)

Sets the read only mode of the ContextGenerator. While in read only mode, only features that previously existed will contribute to context vectors..

Specified by:

setReadOnly in interface ContextGenerator

addContextTerms

protected void addContextTerms(SparseDoubleVector meaning,
                   Queue<String> words,
                   int distance)

Adds the index vector for each co-occurring word in the context. Index vectors are permuted if permFunc is not null. When in read only mode, only existing index vector are used.