Class DistributionalModel

java.lang.Object

DistributionalModel

public class DistributionalModel
extends java.lang.Object

Reads a file collected from Google's Syntactic N-Grams corpus, build an intermediate graphical representation of every line, traverse the graph looking for paths deemed informative by the context model and use those paths to build a tensor of co-occurrence statistics.

Author:

Max Lotstein

Constructor Summary

Constructors

Constructor and Description
DistributionalModel() The constructor of the DistributionalModel

Method Summary

Modifier and Type	Method and Description
void	addToTensor(DependencyPath d, long freq) Accepts a dependency path and an integer representing co-occurrence frequency and updates the tensor object.
java.util.HashSet<DependencyPath>	extract(PartOfSpeechNode[] graph) Given an array of nodes that comprise a graph, use syntactic nGram of arcs, update the tensor to reflect meaningful associations between pairs of content words.
static void	main(java.lang.String[] args) The main loop which opens the NGram file, processes it, and prints the execution time.

Methods inherited from class java.lang.Object

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

Constructor Detail

DistributionalModel

public DistributionalModel()

The constructor of the DistributionalModel

Method Detail

addToTensor

public void addToTensor(DependencyPath d,
                        long freq)

Accepts a dependency path and an integer representing co-occurrence frequency and updates the tensor object. NOTE: in previous versions, the tensor entry would be POS-sensitive but in this version, it is POS-agnostic.

Parameters:

d - a DependencyPath

freq - an integer, representing the number of times the dependency path was observed in the corpus

extract

public java.util.HashSet<DependencyPath> extract(PartOfSpeechNode[] graph)

Given an array of nodes that comprise a graph, use syntactic nGram of arcs, update the tensor to reflect meaningful associations between pairs of content words.

Parameters:

graph - An array of PartOfSpeechNodes that represents a graphical representation of the dependency parse

Returns:

A set of dependency paths

main

public static void main(java.lang.String[] args)

The main loop which opens the NGram file, processes it, and prints the execution time.

Parameters:

args - The name of the NGram file