CoOccurrence

Class FactoryforMatrixSWsynonyms

java.lang.Object

CoOccurrence.FactoryforMatrixSWsynonyms

public abstract class FactoryforMatrixSWsynonyms
extends java.lang.Object

Version:

\date

Author:

Nina Tahmasebi This class provides a framework for creating a dictionary as well as a co-occurrence matrix based on input. Factory provides the framework for creating a dictionary from a term list. Currently the term list is created in Perl and starting with a term, every other line in the list is the frequency of the previous term. The dictionary will be created so that one can distinguish between final terms, which are terms that appear as they are in the term list, and part-of-term which are parts of the multi-word terms from the term list.

Factory also provides a tool for creating co-occurrence matrices using a sliding window w based on a given dictionary. The dictionary can be changed during the course of creating a matrix. The matrix is build in sparse form and is build iteratively. An instance of a co-occurrence matrix must always be included when building a co-occurrence matrix, in the first iteration an empty matrix must be included., Nina Tahmasebi, L3S Research Center

See Also:

CoOccurrenceMatrix, Dictionary

Constructor Summary

Constructors

Constructor and Description
FactoryforMatrixSWsynonyms() Constructor is never used.

Method Summary

Methods

Modifier and Type	Method and Description
static void	addTermsInList(java.util.ArrayList<Term> list, CoOccurrenceMatrix myMatrix)
static Dictionary	buildDictionary(java.io.Reader termList, Dictionary dict) Builds a dictionary based on a term list.
static CoOccurrenceMatrix	buildMatrixGR(Dictionary dictionary, java.io.Reader textInput, CoOccurrenceMatrix startMatrix) Builds a co-occurrence matrix based on a starting matrix, a dictionary and a text file using grammatical relations such as (and, or , comma).
static CoOccurrenceMatrix	buildMatrixSW(Dictionary dictionary, java.io.Reader textInput, CoOccurrenceMatrix startMatrix, int windowsize) Builds a co-occurrence matrix based on a starting matrix, a dictionary and a text file using a sliding window.
static void	convertFiles(java.io.Reader r, java.io.Writer w) Specially written for e-books where text is chopped up in 8-10 words per line.
static CoOccurrenceMatrix	createMatrixFromFile(java.io.Reader matrixSource, Dictionary dict) Creates a CoOccurrenceMatrix from a file.

Methods inherited from class java.lang.Object

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

Constructor Detail

FactoryforMatrixSWsynonyms

public FactoryforMatrixSWsynonyms()

Constructor is never used.

Method Detail

buildDictionary

public static final Dictionary buildDictionary(java.io.Reader termList,
                         Dictionary dict)
                                        throws java.io.IOException

Builds a dictionary based on a term list. All terms in termList are read and added in the hash map in the following way;
"a b c" --> ("a", 1), ("a b", 1), ("a b c", -1)
if "a" is also added, dictionary will have the following input;
(a,2), (a b, 1), (a b c, -1)
Part-of-terms are stored with 1
Final terms are stored with -1
Terms that are both 1 and -1 are stored with 2

Parameters:

termList - Each line should contain a term and term frequency where the term and the frequency are separated by one tab.

Returns:

The dictionary created from termList of the form described above

Throws:

java.io.IOException

createMatrixFromFile

public static final CoOccurrenceMatrix createMatrixFromFile(java.io.Reader matrixSource,
                                      Dictionary dict)
                                                     throws java.io.IOException

Creates a CoOccurrenceMatrix from a file.

Parameters:

matrixSource - the location from which the CoOccurrenceMatrix should be created

dict - the dictionary associated with this CoOccurrenceMatrix

Returns:

A CoOccurrenceMatrix created based on matrixFile

Throws:

java.io.IOException

buildMatrixSW

public static final CoOccurrenceMatrix buildMatrixSW(Dictionary dictionary,
                               java.io.Reader textInput,
                               CoOccurrenceMatrix startMatrix,
                               int windowsize)

Builds a co-occurrence matrix based on a starting matrix, a dictionary and a text file using a sliding window. Returns an extended version of startMatrix based on co-occurrences found in textInput using dictionary and a sliding window of specified size. For first iteration startMatrix is an empty matrix, following iterations startMatrix should the matrix from the previous iteration.

Parameters:

dictionary - use the dictionary that is created from the collection where matrixInputreader comes.

textInput - contains the text from which we should find co-occurrences from the terms in dictionary

startMatrix - matrix representing the previous iteration

windowsize - the size of the window in which we want to find co-occurrences

Returns:

A co-occurrence matrix that is based on startMatrix and extended with the co-occurrences found in matrixInputreader

Throws:

java.io.IOException

convertFiles

public static void convertFiles(java.io.Reader r,
                java.io.Writer w)
                         throws java.io.IOException

Specially written for e-books where text is chopped up in 8-10 words per line. Conversion is done by fusing lines together as long as they are not separated by an empty line.

Parameters:

r - original source of text that needs to be converted

w - source where text should be stored after conversion

Throws:

java.io.IOException

buildMatrixGR

public static final CoOccurrenceMatrix buildMatrixGR(Dictionary dictionary,
                               java.io.Reader textInput,
                               CoOccurrenceMatrix startMatrix)
                                              throws java.io.IOException

Builds a co-occurrence matrix based on a starting matrix, a dictionary and a text file using grammatical relations such as (and, or , comma). Returns an extended version of startMatrix based on co-occurrences found in textInput using dictionary and "and, or, comma" relations. For first iteration startMatrix is an empty matrix, following iterations startMatrix should the matrix from the previous iteration.

Parameters:

dictionary - use the dictionary that is created from the collection where matrixInputreader comes.

textInput - contains the text from which we should find co-occurrences from the terms in dictionary

startMatrix - matrix representing the previous iteration

Returns:

A co-occurrence matrix that is based on startMatrix and extended with the co-occurrences found in matrixInputreader

Throws:

java.io.IOException

addTermsInList

public static void addTermsInList(java.util.ArrayList<Term> list,
                  CoOccurrenceMatrix myMatrix)