Implementation of a Tool for Ontology Paraphrasing, Querying and Visualization on the Semantic Web

schoolmistInternet and Web Development

Oct 22, 2013 (3 years and 9 months ago)

95 views

Implementation of
a Tool for Ontology
Paraphrasing, Querying and
Visualization on the Semantic Web


Goal of the Project:


The
objective of the project is to
develop a
T
ool for
Paraphrasing
,

Querying
and
Visualization
from the ontology

represented as OWL.

The developed
Visualization and Querying Tool would be used by CAIR (DRDO).

Background:


Ontology is defined as an explicit specification of conceptualization. It is a
way of representing things and the relationship between the things in such a way
that a

machine can be made to understand it. OWL (Web Ontology Language) is a
W3C recommended ontology language is OWL for the Semantic Web Activity.


Ontology is the basis of Semantic Web because all web resources and the
relationships between those web resour
ces should be kept track off. So that we can
perform context based searching or perspective based filtering over the web.


Ontology is a mechanized
way of Knowledge Representation. When
Knowledge is represented as ontology then
Inference Engines and Reason
ers

can
be used to mine useful information that can be used to extend the
already
existing
knowledge

about a particular domain
. O
WL

is a good knowledge representation
language.


The goal behind developing ontology
is to use the fact
s and relationship
s

repr
esented in the ontology to extract or infer information that helps user to
perform analysis of the facts. Hence we need tools to Create, Edit, Visualize,
Query, and Paraphrase the ontology.

There are few tools available for doing the
above said actions.


V
isualization of ontology is the process of representing the ontology resources
and their relationships as
a visual
graph. Querying
from ontology
is the process of
extracting us
eful result
s out of the ontology based on certain inputs and constraints.

Para
p
hrasing is the process of building sentences or paragraph out of the fa
cts
extracted from the ontology based on the templates specified
by the user or based
on systemic

functional
grammars.

The steps in paraphrasing are document
planning, micro
-
planning an
d surface realization.


2


Problem Description:


Ontology represented in owl format should be

parsed. The identified ontology
resources

and the relationships after parsing should be mapped to a Graph data
structure. The built Graph data structure should be vi
sualized.
Querying support for
SPARQL/RDQL should be provided to query from the ontology. Paraphrasing
facility should be provided in a custom attributes based

aprroach
.


This tool should be web based because the client should be able to interact
with the
tool through a web browser.
All the facilities should be provided in a
single web base
d

graphical user interface. The tool should be accessible from a
remote client

through local network or Internet.


Following issues should be addressed while developing t
his tool. They are

1)

Tool should be fully browser based

2)

Tool should be available in operational mode in anytime
-
anywhere basis.

3)

Tool should not be based any proprietary libraries or software.


Existing Work & Issues:


There are few softwares available for C
reating, Visualizing,
Paraphrasing,
and
Querying ontology. But m
ost of them are
commercial

and not open source
. They
are the products of several research groups and many
are black
-
boxed ones and not
open source software.


Following are issues that are foun
d in the existing softwares:

1)

None of them is browser based ones.

2)

No full fledged tool available for paraphrasing from the ontology.

3)

Most of them are black
-
boxed softwares, not open source.






3


Proposed Work:


A

web application for pa
raphrasing and queryi
ng from
ontology with
visualization support
for AINN group of CAIR (DRDO) which is fully browser
based.


First one is to build a Graph Constructor

to load and parse the owl file and to
identify the classes, properties and individuals avail
able in the ontol
ogy. The
constructor

lists out the resources found the ontology as the output.

It

should be developed to build graph
by
traversing through the resources and
making the resources as nodes and their relationships as edges. After traversing all
the resources

it produces a OWL Graph as the output.

A
Customized Visualizing Utility

should be built to render the OWL Graph
visually on the screen base
d on the resource types, their relationship types and user
given custom parameters.

A Customized Query Engine should

be built to process the SPARQL/RDQL
query and the results should to be displayed in both visual and textual forms. It
should be capable of checking the validity of the query, process the query and
return results.

A Customized Paraphrasing Utility should b
e built to paraphrase facts
extracted either by querying or based on certain parameters.

It focuses on building
complete sentences from the facts derived.
It is planned to follow template based
approach.

All the Functionalities should be available in
singl
e graphical user interface
and should be of click and use type.


Software Requirements:


1) JDK 1.5 (Java Development Kit)


2) JRE 1.5 or above enabled Web Browser


3
) JUNG 1.7.6 (Java Universal Network/Graph Library)


4
) JENA 2.5.7 (
Java framework for b
uilding Semantic Web applications)




4


Block Diagram:


OWL Graph

Construction



Graph



Paraphraser



Query
Processor


Visualizer


Ontology

OWL
File

owl graph

o
wl graph

Summary

Result

5














Architecture
Diagram:


OWL Graph

Construction



Graph



Paraphraser



Query
Processor


Visualizer


Ontology

OWL
File

owl graph

owl graph

Summary

Result

6




















The user interacts with the system through browser
. The user can access the
tool

from any rem
ote location. The application interacts with the
Web S
erver
and uses
the owl files available in the
web
server.







Modules


OWL Files


Java Library


JUNG,JENA
Library


Ontology
Querying and
Visualization
Applets




Browsers



Java Applet Class File

Web
Server

Client

7



The
Paraphrasing

and Querying tool comprises of
the
following modules.

1
.
OWL Graph Constructor


This module is responsible for r
eading an owl file from the server and
capturing all the resources of ontology such as classes, properties, individuals,
relationships and other elements of ow as whole structured model.

It

converts the
identified classes, propertie, individuals and few o
ntology resources into owl nodes
and the relationships between them as owl edges. Both the owl nodes set and owl
edges set form the owl graph.

2
. Template Based or Seismic Grammar Based Paraphraser


This module helps to paraphrase the facts represented as
ontology based on the
user attributes like start owl node and distance to cover.Paraphrasing is either done by
following Template based approach of Seisimic Grammar based approach.

3.
OWL Graph Visualizer


This module is responsible for setting the layout
of the owl graph, visual
properties of owl graph nodes like owl classes, owl properies, owl individuals like
colour, size, shape, font, icon, etc., and visual properties of owl edges between the
ontology resources like arrow style, thickness, label font,
colour,etc.

4.
SPARQL Query Processor


This module supports query processing of RDQL or SPARQL queries. It
validates a given query from the user and process the query on the ontology
represented as graph and return back either textual result or visual resu
lt based on the
requirement.








Level


0:
Data Flow Diagram

/ Context Diagram

8




























Level
-

1
:

Data Flow Diagram

Paraphrased
Summary

Query Result

Visual Graph







Ontology

Paraphrasing,

Querying
and
Visualization
Tool

OWL

9






























Level
-
2: Data Flow Diagram

OWL
Graph

OWL
Graph

OWL
Graph

Visual
Grap
h

Natural
Language
Summary

Query
Result

OWL

1



OWL
Graph

Builder



3


Query

Processor


Query



User


2




Paraph
r
aser


4




Vis
ualizer


Concept, Distance

OWL Graph

10


For OWL Graph Construct
or

(1):























Input


:

OWL File

Output


:

OWL Graph


Algorithm

:


1) Check whether the file is available or not


2) If available then open the file and load the contents


3) Iterate through the c
ontents of the file and map the i
dentified


ontology resources such as class, property, individual and

other OWL
constraints to OWL Nodes and OWL Edges and Build the OWL Graph.


4) Return the OWL Graph.




Level
-
2: Data Flow Diagram

1.1

Read the OWL File

and Load the
Ontology

1.2

Map the
Ontology
Resources

to
Nodes and Edges

OWL Nodes and
Edges

OWL

Ontology Model

1.3

Build the G
raph

Using the Nodes,
E
dges

and OWL
Co
nstraints.


OWL Graph

11


For Paraphraser(
2
):
























Input


:

OWL Graph, Concept, Distance

Output


:

Natural Language Description


Algorithm

:

1)

Traverse the OWL Graph and get the Sub OWL Graph based on the
distance and concept.

2)

In the Sub OWL Graph perform the required facts search

3)

Apply the Templates/Sei
smic grammar to frame natural language
sentences from the facts.

4)

Return the Natural Language Description



Level
-
2: Data Flow Diagram

2
.1

Select Contents of
Interest

2
.2

Lexicalize Sentences
and Aggregate them

Formed Sentences

OWL Graph

Required Facts List

Concept & Distance

2
.2

Apply grammar to
the formed sentences.

Natual Language
Summary

12


For Query Processor(3):
















Input


:

OWL Graph, Query

Output


:

Query Result


Algorithm

:

1)

Valida
te the query with the regular expression framed for the SPARQL
quey.

2)

If it is valid then Traverse the graph and find the required facts/sub
-
graph

3)

Return the textual result

4)

Render the OWL Sub GraphText







Level
-
2: Data Flow Diagram

3.1


Validate the Query

3.2

Execute the Query by
Traversing the

OWL Graph.

OWL Graph

OWL Graph

Valid Query

Query Result

Query

13


For Visualizer(4):

























Input


:

OWL Graph

Output


:

Visual OWL Graph


Algorithm

:

1)

Create Layout and renderer for the OWL Graph.

2)

Set OWL Class, OWL Property, & OWL Individual visual properties like
color, icon, font
, label position,etc. of the r
enderer.

3)

Set OWL Edge visual properties like thickness, color, label position,etc.
of the renderer.

4)

Render the OWL Graph
.


Related Works:


4.1

Create

Layout and
Create Renderer

Re
nderer

Renderer

4.3

Render the OWL
Graph

Visual
OWL
Graph

OWL Graph

4.2

Set Visual attributes

for OWL Edges,
OWL Nodes,

14


1)

OntoSpere3D is a visualization tool for ontologies

that uses a 3
-
dimensional
space, where information is enriched by

visual cues (as the colour or the size
of visualized entities). The tool aims at tackling visualization issues for
ontology visual models by adopting a dynamic collapsing mechanism and
different views, at different granularities in order to grant a consta
nt
navigability of the rendered model.


2)

ConceptVISTA, an ontology creation and visualization tool developed by

researchers at the GeoVISTA Center, allows users to define and link concepts

and resources pertaining to a conceptual domain. Users can add an
d view

additional information related to concepts and resources. ConceptVISTA
stores

ontologies in the Web Ontology Language (OWL) format and can import
outside

ontologies. ConceptVISTA provides users with a useful set of ontology

management functions.



References:


1) OWL Web Ontology Language Guide:



http://www.w3.org/TR/owl
-
guide

2) SPARQL:


http://www.w3.org/TR/rdf
-
sparql
-
q
uery

3) “Generating Natural Language Descriptions from OWL


Ontologies”
-

Ion Androutso
poulos and Dimitrios Galanis

4) “Speech and Language Processing”, ISBN 81
-
7808
-
594
-
1


-

Daniel Jurafsky and James H.Martin

5) “Systemic Functional Linguistic
s”


-

http://www.isfla.org/Systemics/

6) “Systemic Functional Grammar”


http://minerva.ling.mq.edu.au/resource
/VirtuallLibrary


/Publications/sfg_firststep/SFG%20intro%20New.html

7) “Paraphrasing from ontology”


http://www.mindswap.org/papers/nlpowl.pdf