KYOTO Logo
Knowledge Yielding Ontologies
for Transition-based Organization
  • Increase font size
  • Default font size
  • Decrease font size

Term graph browser

Description

Stand-alone Java application that reads the export of a term database and a wordnet and represents a mind map of terms that occur in a document collection in a TouchGraph browser for navigation.

Owner: VU University Amsterdam

License: GPLv3

Binaries:

http://kyoto.let.vu.nl/~kyoto/files/kyototermgraph/kyototermgraph.1.0.zip

Source code:

https://kyoto.let.vu.nl/svn/kyoto/trunk/tools/termgraph

https://kyoto.let.vu.nl/svn/kyoto/trunk/tools/touchgraphlayout

Documentation:

External resources:

The Term browser uses the TouchGraph library to display and navigate through the data. TouchGraph is available through the Apache Software License and can be downloaded from: http://sourceforge.net/projects/touchgraph/


Introduction

This package contains classes to read exports from the Kyoto term database and provides various interfaces for displaying them. The data from the term export is stored in the Term class. Various other structures are derived to make it easier to navigate relations.

The program can also read a wordnet in wordnet-lmf format and a list of wordnet synsets to base concepts mappings (BCs). The synsets are stored in the WordnetLexicon class, which has a hash-map from synset Ids to synsets and a hash-map from synset Ids to wordnet-entries. A WordnetEntry has a list of senses.

Wordnet synsets that are referred to in the term export are also stored as terms. If wordnet and BCs are read, the term hierarchy is extended with the wordnet hierarchy on the basis of the synsets that are associated with the terms. If more than one synset is associated, it takes the first highest scoring synset. If no BCs are read, the full wordnet hierarchy is built on top of the term hierarchy. If BCs are read, the tree is cut at the level of the BCs.

There are a several main classes:

TermTree

Main class: eu.kyotoproject.termtree.TermTree

 It generates Html files with all the terms in a hierarchical structure that can be used to classify the terms, split over part-of-speech. It reads term export+wordnet-lmf and outputs html files with the complete hierarchy.

Arguments:

term-export file

wordnet file in Wordnet-LMF format

two-letter language code

TermBC
Main class: eu.kyotoproject.termtree.TermBC

 It reads term export+wordnet-lmf+bc-mapping and outputs html files with the hierarchy up to the BC level.

 Arguments

term-export file

wordnet file in Wordnet-LMF format

list with synset to base concept relations

 two-letter language code
GLTermPanel

Reads a term export+wordnet-lmf (optional)+BC mapping (optional) and display a TouchGraph Panel. You can also load a file from the interface

Main class: eu.kyotoproject.termtree.GLTermPanel
Arguments:

--term-export

--wn-lmf (optional)

--language en/en

The package includes examples of terms exports, English Wordnet in Wordnet-LMF format and shell scripts that illustrate the calls.

Below we show an example of how the browser looks like. It has two panels. In the left panel you can load a term database export file. If you combine it with a wordnet, as is shown, the tree is augmented with wordnet hierarchies. The boxes with black font are new terms that occur in the text but not in wordnet, blue font terms occur in the documents and in wordnet and red font boxes are hypernyms in wordnet that do not occur in the document. You see here the structure acquired for "bird". You can navigate the graph using the scroll bars. Use level to expand more or less levels, and confidence to restrict to terms with lower or higher scores. Next to the load file button you see a drop down list with the terms that are most salient. The right pane is used to display text occurrences of a term. Right-click on a term to see the text occurrences. It makes a call to a remote term database and only works if the term database is online.

 Term graph interface

 

 

 

 

ICT-211423 - 2008 © Kyoto Consortium