=Paper=
{{Paper
|id=Vol-2205/tutorial2
|storemode=property
|title=None
|pdfUrl=https://ceur-ws.org/Vol-2205/tutorial2.pdf
|volume=Vol-2205
}}
==None==
https://ceur-ws.org/Vol-2205/tutorial2.pdf
Conceptual Ontology Engineering Tutorial
Michael G. BENNETT
mbennett@hypercube.co.uk
Hypercube Limited, London, England
Abstract: Conceptual modeling as defined within the discipline of software
development is the exercise of creating computationally independent model
artifacts against which to develop and validate logical and physical model design
artifacts. The art of conceptual modeling is one that requires a clear understanding
of the notion of a concept and an appreciation of the nature of concepts as distinct
from words, labels or database element names. One powerful type of conceptual
model is the ‘ontology’ where ontology is understood to be a formal specification
of a conceptualization. The word ‘ontology’ is broadly used to cover a number of
such specifications. The goal of this tutorial is to present a formal framework
within which to understand these distinctions and to introduce techniques by
which attendees may be able to develop ontologies that may serve as conceptual
models, focusing on the less technical (and often overlooked) aspects of such
ontology development, specifically the ability to appreciate concepts and to model
these within the logical formalisms used in ontology development.
Keywords: Ontology, conceptual model, development lifecycle, concept,
knowledge representation, top level ontology.
1.Introduction
This tutorial sets out the basic principles of concept modeling, situating these kinds of
models within a broader modeling framework that includes logical models, ontologies
for reasoning applications and so on. Attendees will learn how to frame this kind of
ontology artifact, how to think in terms of concepts and how to define these in formal
logic. The bulk of this tutorial focuses on conceptual issues: understanding concepts,
classification theory and powertypes, the use of formal logics in ontology development
and issues relating to terminology and vocabulary. Attendees will learn ontology
development techniques such as the use of upper ontologies to provide disambiguation
of similar concepts and how these abstractions address common data problems.
Specific examples of upper and cross-domain ontologies are covered in depth,
including contextually defined concepts (roles etc.), event and process modeling,
contracts and transactions. The course concludes by identifying the range of ways in
which conceptual ontologies may be used in various practical deployment architectures
and how to use or extend popular ontologies such as the Financial Industry business
Ontology (FIBO), with examples. No prior knowledge of ontology modeling is
required.
This tutorial is intended to describe conceptual ontology modeling techniques and
concerns. It does not cover the technical details of modeling ontologies in RDF and
OWL but is intended to provide the groundwork for such work. Those who are
interested in implementing RDF and OWL models for applications may wish to
complement the learning from this tutorial with a comprehensive course in OWL based
modeling.
1.1.Intended Outcomes
By the end of this tutorial attendees should be able to create their own conceptual
ontologies and understand how the use of ontologies as conceptual models can enhance
software development and cut integration costs. Attendees will also understand how to
derive technical artifacts from these for data integration and model driven
development, as well as pragmatic, operational ontologies for semantically enabled
reporting and inference processing applications.
1.2.Intended Audience
This tutorial is aimed at anyone interested in data modeling, knowledge representation
and systems development, including students, researchers, data architects and business
analysts. It is relevant to anyone exploring the use of formal ontologies for a range of
different application areas, particularly in emerging technology areas such as micro-
finance, distributed ledger technology (AKA Blockchain), big data, machine learning
and the Internet of Things (IoT). No prior knowledge of ontology modeling or
standards is required. Some basic knowledge of information technology is assumed,
including familiarity with the technology development lifecycle, but no prior
knowledge of any language for programming, databases or modeling is assumed.
2.Outline of the Tutorial
Introduction: Concepts and words; the data development lifecycle; use of
computationally independent models. Introducing ontology: a conceptual model
for data and beyond.
Modeling Semantics: principles of semantic modeling, illustrated with a rolling
example. Defining concepts. Classification and taxonomy, properties, the
differentiating characteristics of concepts; understanding formal logic and set
theory. Formal semantics basics – representation of classes, properties and logical
restrictions.
Conceptual Issues: Anatomy of a Concept; words, concepts and lexical space.
Homonyms, heteronyms and some strange habits of words. Concepts without
words. Different approaches to formal semantics.
Classification principles: kinds of taxonomy; subsumption based taxonomies; faceted
classification; powertypes and kinds of individuals.
Introducing Data: Distinguishing things from data about things; semantic ‘truth-
makers’ versus data; real things that are data; establishing data applicability
(semantic distance) for a given type of ontology; datatype properties in ontologies;
information kinds and the use of a ‘values’ ontology.
Top Level Ontologies (TLOs) and Cross Domain Ontologies: Why top level
ontology? Understanding existing top level ontologies and standards. Semantic
abstraction and re-use; dimensions of a top level ontology. Some popular top level
ontologies. Things defined by their context; things that happen; other partitioning
considerations. Realism versus concept-centric ontology.
TLOs In Depth: Contextual Things Deep dive session on things in roles and other
contextual matter. Different conceptualization options for roles and relators.
Examples of these using customer and counterparty data modeling issues.
TLOs In Depth: Things that Happen Deep dive session on continuant and occurrent
things (endurants and perdurants). Classifying kinds of occurrent. Different
conceptualization options for things that ought to happen or might happen. The
semantics of plans and risks. Modeling business processes as ontology.
Recommended Mid-level Ontologies: Authoritative Sources of Meaning: identifying
meaningful published concept definitions and adapting these into the ontology
framework using TLO (with examples). The REA ontology for transactions; LKIF
and other legal ontologies; ontologies for business process and other common
problem areas.
Conceptual Ontology Development: Framing ‘Simplest kind of Thing’ concepts
(archetypes); top down, bottom up and middle out ontology development; the use
of the ‘wire frame’ upper ontology for pragmatic conceptual ontology
development.
Putting it to Work: Business concept ontologies versus application ontologies.
Introducing the Financial Industry Business Ontology (FIBOTM) Standard. Putting
these to use: mapping, reporting, inference processing, Blockchain, graph
analytics, machine learning, legal and regulatory (RegTech) and novel finance and
micro-finance opportunities (FinTech). Styles of ontology for different ontology
uses (with examples). Getting to there: a roadmap for ontological maturity.