Practical Knowledge Modelling: Ontology Development 101

This lecture talks about the different types of knowledge and how knowledge is consumed by us. We'll touch on the generic context of knowledge modelling and see how it falls in the much wider area of knowledge management and information systems.

We very often assume we know what the differences are between data, information and knowledge. This lecture will clarify what we mean by these three concepts in the context of knowledge modelling.

This lecture will give you a taster of what a knowledge model (a.k.a. ontology) is. We'll look, conceptually and at a very high level, what a knowledge model consists of by drawing some useful analogies and without going too technical.

There's a much more profound underpinning when we talk about ontologies. This lecture looks at the philosophical perspective of ontologies and introduces the idea of different levels of abstraction for making sense of the entities in the world around us.

In this lecture, we'll take a look at the most fundamental components of a knowledge model. We'll introduce the concepts of classes, relationships, individuals and axioms that we can use to describe a particular subject matter.

The representation of knowledge models can be tailored for human and machine interpretation. In this lecture, we'll run through the basics of what's needed for being able to represent ontologies.

In addition to the philosophical perspective, ontologies can also be viewed from a logic based perspective. This lecture covers, at a relatively high level, what the logic based perspective is about.

In this lecture, we'll elaborate a little more on the different levels of abstraction in knowledge modelling. We'll take a look at how the stack of abstraction is layered, from the generic to the more specific.

Knowledge modelling has a great many applications, from things like knowledge management to Artificial Intelligence. This lecture summarizes what these applications are.

In this lecture we will see some concrete real-world examples of applied ontology.

This lecture concludes Section 2, summarising the main points discussed.

To conduct knowledge modelling the best way possible, you need a methodology that helps you manage your knowledge modelling expectations. This lecture introduces a methodology that brings out the best of a simple approach but providing sufficient coverage for the full life cycle of knowledge models.

Requirements manage is a concept used in things like project management and business analysis. It can also be used successfully in knowledge modelling. So, in this lecture, we'll take a look at why we need to manage requirements and how we go about doing that.

This lecture explains the goal and scope definition phase of our methodology. This is the phase where you highlight the high level aims, objectives, scope, etc., of your knowledge modelling activity. In this course, we'll be looking at how we go about building a knowledge model of different types of ballpoint pens based on their components and characteristics. We'll be using the ballpoint pen example throughout the course.

Some ontology development methodologies identify “competency questions” as a means of scoping a knowledge model. This lecture elaborates on this topic.

When modelling knowledge, you very often start off gathering unstructured or semi-structured information that you then later analyse. This lecture talks about the methods for gathering and eliciting information, as well as their advantages and potential drawbacks when it comes to collecting information for knowledge modelling.

In this lecture, we'll focus on explaining a couple of really useful methods, namely affinity diagrams and mind maps, to help us quickly spot early patterns in the organisation of ideas and concepts.

This lecture concludes Section 3, summarising the main points discussed.

In this lecture, we'll introduce the initial structuring phase of our knowledge modelling methodology. Initial structuring is about transforming the unstructured or semi-structured information you collected in the previous phase into visually-represented knowledge models.

How can go from unstructured or semi-structured information to structured models? The starting point is about being able to analyse all sorts of content. This lecture looks at how we go about listing and analysing statements to start spotting ontology structures.

The analysis of statements leads to the compilation of a term pool, which you need to administer during the course of ontology modelling. We'll explore an approach, based on a spreadsheet to help you with easily tracking and monitoring the progress of your terms, as you make decisions as to what needs to go into your knowledge models and what you leave behind.

This lecture introduces the implications of choosing graphical languages for modelling the building blocks of knowledge models.

Unified Modelling Language (UML) is just one graphical method for representing subject matter. In this lecture, we'll see what UML has to offer to help us model ontology structures.

This is a continuation from the previous lecture. This time, we'll be looking at actually using a software application for representing knowledge models expressed in UML. If you do not have a suitable software installed for modelling in UML - no worries, you can simply use pen and paper!

In this lecture, we'll be looking at a second graphical language for representing ontologies, called IDEF5. We'll first introduce the basic shapes we need for modelling and then, we'll dive into doing some modelling work using Microsoft Visio. If you don't have Visio installed, you can still crack it out with pen and paper!

In this lecture, we'll be going through additional discussions on the topic of capturing and representing knowledge using graphical methods.

This lecture concludes Section 4, summarising the main points discussed.

The formalization phase of our methodology is about transforming our visual knowledge models into models that can be interpreted by computers. In other words, we'll learn how to code our models but without needing to do any coding at all, as a user interface application will handle that for us! We'll be going through an in-depth explanation of the Web Ontology Ontology (OWL) and see how to use Protégé ontology editor. In this section, we'll be showing how OWL works by going through various hands-on tutorials based around our ballpoint pen ontology.

In this lecture, we'll download Protégé ontology editor and also run through the steps for setting up an ontology in Protégé.

In this activity, you will download and run Protégé ontology editor, as well as provide an IRI for your ontology.

This lecture covers the basics of creating classes in OWL using Protégé.

This is a continuation from the previous lecture, where we'll look at further features of OWL classes.

In this activity, you will create a hierarchy of classes and declare disjoint classes.

This lecture covers the basics of creating individuals in OWL using Protégé.

In this activity, you will create individuals of classes.

This lecture covers the basics of creating properties in OWL using Protégé.

The domain and range allow you to specify the classes or sets of classes that your relationships are expected to connect to. In this lecture, we'll focus on how to specify the domain and range.

This lecture explains how to create inverse properties in OWL using Protégé.

Property characteristics help us build more meaningful relations in our knowledge models. This lecture explains all the property characteristics supported in OWL.

In this activity, you will create a hierarchy of object properties, define the domain & range of properties, define inverse properties and specify property characteristics.

By describing and defining classes, it becomes possible to build more meaningful class structures. In this lecture, we'll cover the basics of class description and definition in OWL.

Existential restrictions are one kind of description we can add to classes to capture their semantics. This lecture explains and exemplifies how to use existential restrictions in OWL.

In this activity, you will create existential restrictions.

'hasValue' restrictions are another kind of description we can add to classes to capture their semantics. This lecture explains and exemplifies how to use 'hasValue' restrictions in OWL.

In this activity, you will create hasValue restrictions.

This lecture explores the essence of primitive and defined classes in OWL.

In this activity, you will convert primitive classes to defined classes.

Ontology tools for OWL offer reasoning facilities for making inferences based on the structures and descriptions captured in our formal knowledge models. In this lecture, we'll see how to use a reasoner in Protégé to perform auto-classification of concepts for us.

In this activity, you will work with the inferred class hierarchy view and invoke a reasoner to reclassify the classes.

Universal restrictions are another kind of description we can add to classes to capture their semantics. This lecture explains and exemplifies how to use universal restrictions in OWL.

In this activity, you will create universal restrictions.

Cardinality restrictions are another kind of description we can add to classes to capture their semantics. This lecture explains and exemplifies how to use cardinality restrictions in OWL.

In this activity, you will create cardinality restrictions.

In OWL, datatype properties are relationships that link something to a data value. This lecture exemplifies how to work with datatype properties.

In this activity, you will create and work with datatype properties.

Class description and definition do not stop there. There are other ways in which we can capture the description of our classes. In this lecture, we'll be swinging by these extras.

In this activity, you will create a class as a complement of another.

Protégé comes with various plugins and tools for helping us do a range of things like visualizing our knowledge models, while we create them in the application. In this lecture, we'll run through the other useful features you need to be aware of.

In this activity, you will explore other useful features in ontology formalization.

This lecture concludes Section 5, summarising the main points discussed.

The deployment phase of the methodology is for us to exploit our knowledge models in their intended application or settings. This lecture introduces the idea of deployment, highlighting the considerations to bear in mind when rolling out our ontologies.

This lecture looks at a few tools and methods to generate ontology documentation from OWL files.

In this lecture, we'll be focusing on additional ways to share our knowledge models using visual and graphical methods.

This is a continuation from the previous lecture, where we'll look at the use of radial diagrams for composing ontology visuals.

This lecture explores, at a conceptual level, what the building blocks of ontology driven systems are. We'll discuss the basic architecture for being able to 'plug' our formal knowledge models into actual information systems for people to start using.

This lecture concludes Section 6, summarising the main points discussed.

The evaluation phase is the last one in our methodology. This phase looks at all the previous stages and how well we accomplished them. In this phase, you also try to understand how to continuously improve the whole process of ontology development and deployment, in light of future knowledge modelling cycles.

Lessons learnt is a knowledge management concept that we can apply during the evaluation phase of the knowledge modelling methodology. Here, we'll touch on the basics of lessons learnt and its importance.

This lecture deals with the commonly made mistakes during ontology development that could affect the quality of your work.

This lecture concludes Section 7, summarising the main points discussed.

This lecture provides an overview of the 'softer' skills to target when learning to become a great knowledge architect.

This is the concluding lecture in this series. The lecture briefly wraps up everything we've covered in the course.

Download course slides.

Attributions, special thanks and disclaimer.