from text and ontology : methodologies and tools - Text2Onto

What is O/ontology ? ontology life-cycle Why the need of methodologies Ontology Learning Text2Onto Scientiﬁc research n
From text and ontology : methodologies and
tools
Radhouene ROUACHED
Master Information System Techniques
National Engineering School of Tunis
January 20, 2017
1 / 35

Outline
1 What is O/ontology ?
2 ontology life-cycle
3 Why the need of methodologies
UPON
CommonKADS
Ontology Development 101
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
Modeling Primitives Library
Data-driven change discovery
6 Scientiﬁc research news
2 / 35

« the science of being qua being »[6]
Understanding the reality through the structure of things and
their nature and goes further more in smashing the limits by
studying ﬁctious entities.
« An explicit speciﬁcation of conceptualization »[4]
Applying the mathematical logic in creating new ontologies as
computational models for automatizing reasoning
3 / 35

Ontology life-cycle is deﬁned as a speciﬁc sequence of activi-
ties in order to develop an ontology starting from identifying the
purpose until the assessment phase.
Identify purpose
Acquiring knowledge
Building ontology
Conceptualization
Integrating
Implementation
Documentation
Evaluation
4 / 35

Methodologies
Methodologies are made to insure a good sequence in develop-
ing and maintaining an ontology.Therefore Methodology should
structure the idea from its inspiration until its development and
real implementation in our real day-life and in this way humanity
progresses.
5 / 35

UPON
Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
6 / 35

UPON
UPON stands for Unified Process ONtology.It takes a use case
driven, iterative and incremental nature.The development pro-
cess is a work-flow cycle.It contains 5 main phases[1] and in
each phase reside several steps as { inception, elaboration, con-
struction and transition }:
1 Defining Requirements
2 Analysis
3 Design
4 Implementation
5 Test
7 / 35

UPON
8 / 35

CommonKADS
Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
9 / 35

CommonKADS
CommonKADS (Knowledge Analysis Development System) is
a methodology based on three main aspects, knowledge man-
agement, knowledge analysis and system development1. this
methodology aims to build a knowledge model.
1 Knowledge management
Provides a strong tool for supporting knowledge
management.
2 Knowledge analysis
Aims for the conceptual level through modeling processes
by templates as a result it allows a reusable knowledge.
3 Knowledge system development
Plays a key role in running specifications which firstly
defined to use.
10 / 35

CommonKADS
Supported Tool
PCPACK integrates suite of 10 knowledge tools designed to support
the acquisition and use of knowledge2
.
11 / 35

Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
12 / 35

Ontology Development 101 is a simple knowledge-engineering
methodology in 7 steps .It presents a guide for creating ontolo-
gies.The steps are :
1 Identifying domain
2 re-use of existing ontologies
3 Enumerate important ontologies
4 Identifying classes
5 Identifying properties
6 Identifying facets
7 creating instances from classes
13 / 35

Supported Tool
One of the most well-known ontology development environment
that uses this methodology is protégé. Protégé’s plug-in ar-
chitecture can be adapted to build both simple and complex
ontology-based applications. Developers can integrate the out-
put of Protégé with rule systems or other problem solvers to con-
struct a wide range of intelligent systems.3
14 / 35

DOGMA
Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
15 / 35

DOGMA
DOGMA (Developing Ontology Grounded Methods and Applica-
tions)
Constructing a a guideline for “ontology builders towards building on-
tologies that are both highly reusable and usable, easier to build,and
smoother to maintain” (Jarrar and Meersman, 2007).As a framework,
it combines the relation between axiomatization domain and axiomati-
zation application.
16 / 35

DOGMA
1 Preparatory
1 Definition purpose
2 Usefulness study
3 Preparation and scoping
2 Domain conceptualization
1 Knowledge elicitation
2 Knowledge breakdown
3 Knowledge negotiation
4 Knowledge discovery
3 Application specification
1 Structuring
2 Definition of competency questions
3 Definition of semantic constraints
4 Answering of competency questions
17 / 35

DOGMA
Supported Tool
DOGMA Studio Workbench is the tool suite behind the DOGMA on-
tology engineering approach. It contains both a Workbench and a
Server. The Workbench is constructed according to the plug-in ar-
chitecture in Eclipse. There, plug-ins, being loosely coupled ontology
viewing, querying or editing modules support the different ontology
engineering activities and new plug-ins continuously emerge.4
18 / 35

Ontology Learning
Ontology engineer interacts with the ontology management
component. This last contains the Graphic User Interface (GUI)
and the ontology management back-end.Its main aim is to
make a middle-ware between the ontology and the learning
algorithms. The essential functions for Ontology Learning are:
Evolution
Reasoning
Evaluation
19 / 35

20 / 35

Text2Onto
Text2Onto is an ontology learning framework based Java to support
the acquisition of ontologies from textual documents.5
GATE
GATE (General Architecture for Text Engineering) open source soft-
ware capable of solving almost any text processing problem.6
WordNet
WordNet is a large lexical database of English. Nouns, verbs, adjec-
tives and adverbs are grouped into sets of cognitive synonyms (synsets).7
21 / 35

Install & Conﬁgure Text2Onto
22 / 35

Architecture
Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
23 / 35

Architecture
24 / 35

Architecture
GUI (Graphical User Interface) : Is composed of different
views for the conﬁguration of the ontology learning process and
the presentation of the results.
MPL ( Modeling Primitives Library) : Contains the modelling
primitives in a declarative way.
Ontology Writer : Translate instantiated modelling primitives
into a speciﬁc knowledge representation language.
POM (Probabilistic Ontology Model): Store the result of
different ontology learning algorithms.
Reference Manager : Control and access database.
Algorithms : Algorithms for ontology learning.
Algorithm Controller: Initialize Algorithms , trigger a linguistic
processing(NPL) of data in corpus and Finally apply changes in
the POM.
NPL (Natural Language Processing) : Support curation of a
chemical ontology.
25 / 35

Algorithms
Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
26 / 35

Algorithms
The execution of each algorithm consists in three main phases:
Notiﬁcation phase
The algorithm make an up-to-date about the changes made over
the corpus
Computation phase
Changes are mapped with their references.Then stores these
knowledge relation about ontology and data.Otherways saving
pointers to all occurrences of a concept.
Result phase
Requests for the POM changes are generated from the update
contents of the reference repository.
27 / 35

Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
28 / 35

Modeling primitives used by Text2Onto are :
Concepts (CLASS)
Concept inheritance (SUBCLASS-OF)
Concept instantiation (INSTANCE-OF)
Properties & relations (RELATION)
Domain & range restrictions (DOMAIN & RANGE )
Mereological relations (PART-OF)
Equivalence
29 / 35

Outline
UPON
CommonKADS
DOGMA
4 Ontology Learning
5 Text2Onto
Architecture
Algorithms
30 / 35

Change capturing
It is the generation of ontology changes from explicit
and implicit requirements.
Change discovery
Its main goal is to generate implicit requirements through
activating ontology changes from existing data.
Structure-driven
Structure-driven changes are the deduction from the
ontology structure.
Usage-driven
Usage-driven changes are resulted from the usage
patterns depending on a period of time
Data-driven
Ontology Model Knowledge is changing followed by
methods provides by data-driven which itself depends on
modiﬁcation applied to data set.
31 / 35

Scientiﬁc research news
‘[D. Zhang et al.,2016]’ A new cognitive model for
autonomous ontology learning
‘[B. El Idrissi et al.,2016]’ Supporting collaborative work on
ontology learning from Relational Databases
‘[R. Rupasingha et al.,2016]’ Calculating web service
similarity using ontology learning with machine learning
‘[A. Kamal et al.,2016]’ OntoLSA—An Integrated Text Mining
System for Ontology Learning and Sentiment Analysis
32 / 35

Conclusion
Ontology Learning is the old new era of developing ontologies.It
is linked with many computer science ﬁelds and it is all about
understanding the reality through the structure of things.
33 / 35

Núria Casellas.
Legal Ontology Engineering, volume 3 of Law, Governance and Technology
Series.
Springer Netherlands, 2011.
Philipp cimiano et al.
Handbook on Ontologies, chapter Ontology Learning, pages pp 245–267.
Springer Berlin Heidelberg, 2009.
Nathan Scott Davis.
An Analysis of Document Retrieval and Clustering Using an Effective Semantic
Distance Measure.
Thesis, Brigham Young University - Provo, 2008.
Thomas Gruber.
A translation approach to portable ontology speciﬁcations.
pages 199–220, 1993.
Johanna Völker Philipp Cimiano.
Natural Language Processing and Information Systems, chapter Text2Onto,
pages pp 227–238.
Springer Berlin Heidelberg, 2005.
Christopher Shields.
Being Qua Being , The Oxford Handbook of Aristotle.
2012. 34 / 35

Thank You for your attention.
35 / 35

from text and ontology : methodologies and tools - Text2Onto

More Related Content

Viewers also liked (20)

Similar to from text and ontology : methodologies and tools - Text2Onto (20)

More from RadhoueneRouached (7)

Recently uploaded (20)

from text and ontology : methodologies and tools - Text2Onto