Case Study: Ontology for Metadata in e-Learning

Ana María Fermoso García

, Salvador Sánchez Alonso

Encarnación Beato Gutierrez

and Roberto Berjón Gallinas

Computer Sciences Faculty, Pontifical University of Salamanca, Compañía 5, Salamanca, Spain

Information Engineering Research Unit, University of Alcala, Alcalá de Henares, Spain

Keywords: Learning Object Metadata, Learning Object, IEEE LOM, OWL, e-Learning, Organic.Edunet.

Abstract: Learning objects are a fundamental concept in new educational paradigms based on e-learning. These

resources are not only content but also include metadata descriptions organized into categories, often in the

form of an application profile of the IEEE LOM standard. These metadata descriptions facilitate the design,

search, exchange and reuse of educational contents reachable through learning object repositories. The aim

of this paper is to present the case of the Organic.Edunet portal, where the semantic search capabilities

implemented made necessary to design an OWL mapping for the IEEE LOM standard. We called this effort

LOM2OWL. LOM2OWL is in fact an ontology aimed at improving the search, reuse and use of learning

objects stored in the Organic.Edunet Web portal and repository.

1 INTRODUCTION

Organic.Edunet is a learning portal that provides

access to digital learning resources on Organic

agriculture and agroecology and aims to facilitate

access, usage and exploitation of such content. The

result of the European project Organic.Edunet, it

makes use of IEEE LOM metadata standard for the

educational annotation of resources and the aim is

enhancing its semantic capabilities. Given that

Organic.Edunet provides different mechanisms for

finding resources based on computational

semantics, it was necessary to define a mapping

from the human-oriented IEEE LOM metadata

records that content contributors provided to a

semantic version we received To attain this aim, we

suggested to adapt IEEE LOM to a semantic

language to describe learning object metadata

records in OWL.

This paper reports on this effort and provides as

the most interesting result a series of tables showing

how the mapping was made and the way we

approached to map the different categories in the

standard.

In the semantic Web model the ontologies are

the main tool to present and structure the

knowledge. The most extended language of

ontologies is OWL (Ontology Web Language)

(W3C-OWL, 2011). We think that a LOM to OWL

mapping would facilitate the creation and

management of learning objects metadata by

automated engines making use of Seemantic Web

capabilities, easing at the same time their storage in

semantic repositories such as Ont-Space (http://

code.google.com/p/ont-space/), a Java-based

software framework providing the services of a

semantic metadata repository.

2 IEEE LOM

IEEE LOM is a metadata standard for learning

objects developed by IEEE for the description of

educational resources in digital format through

metadata records. These metadata capture the

essential characteristics of the resource and

organize them according to a number of previously

agreed and established categories.

IEEE LOM establishes a conceptual data

schema that defines the structure of a metadata

record for a learning object. Thus, the first step to

an OWL mapping was to understand these

characteristics, those which are grouped into 9

categories in IEEE LOM. Each category has

subcategories. For example General category has

the following subcategories: Identifier, Title,

Language, Description, Keyword, Coverage,

Structure and Aggregation level.

A brief description of the categories in the

standard would be the following:

317

Fermoso García A., Sánchez Alonso S., Beato Gutierrez E. and Berjón Gallinas R..

Case Study: Ontology for Metadata in e-Learning.

DOI: 10.5220/0004087103170320

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2012), pages 317-320

ISBN: 978-989-8565-30-3

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

1. General: includes general information that

describes the learning object as a whole.

2. Lifecycle: it describes the history and current

state of a learning object, as well as and

those entities that have affected the learning

object during its evolution.

3. Meta-Metadata: it describes the metadata

record itself, rather than the learning object

that a given metadata record describes.

4. Technical: it groups the technical

requirements and characteristics of the

learning object.

5. Educational use: it groups educational and

pedagogical characteristics of the object.

6. Rights: describes the intellectual property

rights and conditions of use of the object.

7. Relationship: defines the relation between a

learning object and other learning objects to

which it is related.

8. Annotation: comments on the educational

use of the learning object and information on

when and by whom these comments were

created.

9. Classification: it describes where the

learning object falls with a particular

classification system (taxonomy).

Besides these categories, LOM also defines the

following data types: LangString for strings,

DateTime for dates, Duration for time periods and

Vocabulary for enumerated types.

Using this schema allows the authors of learning

objects to specify what elements make up a body of

metadata, to facilitate search, evaluation,

acquisition and use of learning objects by students,

instructors or automated systems. In addition it also

facilitates sharing them, allowing the development

of catalogues and repositories.

The inclusion of instances of metadata with the

learning object provides standard information on

the contexts of use, thus increasing their reusability.

Usually this metadata structure is implemented in

XML format. Our effort targeted to map this

structure to an ontology format in OWL.

3 LOM2OWL

As we have seen in the preceding section, learning

objects are characterized by metadata records, each

composed by a set of properties. We thought that

these properties could be used to describe instances

of the learning object from an ontological point of

view. To create an instance of an object implies

first, to define an identifier for the object in order to

identify it, and then to associate values to each

property of the object.

In IEEE LOM conformant metadata records, an

object is described by using nine categories, 1.

General, 2. Life cycle, etc. Each category is formed

by a set of related properties called subcategories.

For example the General category has the

subcategories 1.1. Identifier, 1.2. Title, 1.3.

Description, and so on. Some of these categories

have recursively new subcategories, such as for

instance the subcategory 1.1. Identification, that is

described using characteristics 1.1.1. Catalogue and

1.1.2. Entry. This hierarchical structure of

categories and properties to define an object made

easier its translation to an ontological schema.

In the following sections we will present the

classes defined in the LOM2OWL ontology which

correspond to the knowledge in IEEE LOM

metadata elements. These classes will be used not

only to represent the LOM data types, but also to

describe any LOM record.

3.1 Mapping IEEE LOM Data Types

to LOM2OWL

To represent the LOM data types in OWL we have

defined one class per data type. Table 1 shows the

correspondences.

Table 1: Correspondence between LOM data types and

classes of LOM2OWL.

LOM

Datatype

OWL Class

DateTime

dateTime has the subclass lomDateTime

which has two properties:

textDescriptor and timeItem

LangString

langString specifies the language of a

string. It is composed by various instantes of

the simpleLangString class

Duration

lomDuration has two properties, one to

define the duration and other for its

description, similat to IEEE LOM duration

As it is clear in the table, every data type groups

some characteristics called properties of the data

type. For example the data type LangString has two

characteristics associated: the language of the string

and the string of characters itself.

Another important data type in IEEE LOM is

the data type Vocabulary, which does not have an

equivalent type in OWL. When a property is of this

type, we suggest using a data property of string type

restricted to take values from a fixed list of values

including of course those terms permitted in each

specific vocabulary.

There are several properties of this type. For

example in the 5. Educational category, the 5.5.

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

318

Intended End User Role subcategory means the

principal user(s) with this learning object was

designed to. This feature is encoded in LOM2OWL

using the

isIntendedForUserRole

property of

the

learningObject

. The value space of this

property, which would form its vocabulary are:

teacher, author, learner and manager.

3.2 Mapping LOM Categories

In order to define a metadata record of an IEEE

LOM learning object in OWL, it is first necessary

to define the basic classes which will show the

categories schema of LOM. The main class of the

OWL ontology is

learningObject

which

represents, obviously, the concept of learning

object. Apart from this basic one, we created one

class for each IEEE LOM category. These

categories are described in turn by several

subcategories with grouped element

. Table 2

shows a summary of the main classes of our

LOM2OWL ontology and the equivalent categories

and subcategories in IEEE LOM. At the same time,

Figure 1 shows the hierarchical organization of

these classes.

Each class in the ontology has one or more

properties, one for each of its features. These

properties are usually named by using the

convention

has<property_name>

. As usual in all

ontologies, all properties are defined by a domain

(objects to apply it) and a range (values which the

property can take). Properties can either be of type

objectProperty or dataType.

Figure 1: Classes hierarchy in the LOM2OWL ontology.

Table 2 shows some of the LOM2OWL classes

only, because a full detail of all of them would take

not display very well in the format of this paper. In

any case, we will define in detail the

implementation of one category, for example the

Description of the 1.4.General. Description

category because other classes follow a similar

pattern and can thus be deducted from this example.

This category is aimed at including a textual

description of the content of the learning object. In

LOM2OWL this feature will be implemented using

the property

hasDescription

of the

learningObject

class. Its domain will be the

learningObject

class. Its range will be of

langString

data type. This data type allows more

than one description of the learning object where

each

singleLangString

would provide a

description in a particular language.

4 CONCLUSIONS

IEEE LOM is a model to describe learning object

based on a structure schema hierarchically

organized. This organization make easier to adapt

this model to an ontology with classes and

properties, that in our case we have used to enhance

and promote semantic capabilities in the

Organic.Edunet web portal.

This ontology is currently undergoing a process

of evolution, refinement and polishing as part of the

tasks and activities of the Organic.Lingua project.

ACKNOWLEDGEMENTS

This work is partially funded by the Organic.Lingua

project, EU program CIP-ICT-PSP.2010.6.2, GA:

270999.

REFERENCES

IEEE Learning Standards Committee, 2002: Draft

standard for learning object metadata. In

http://ltsc.ieee.org/wg12/files/. Last access on June

2012.

W3C: OWL Web Ontology Language, 2004. In

http://www.w3.org/TR/owl-features/. Last access on

June 2012.

CaseStudy:OntologyforMetadataine-Learning

319

Table 2: Correspondence between LOM2OWL classes and properties of IEEE LOM categories

LOM2OWL class Equivalent LOM category Properties of LOM2OWL class

contribution

learningObjectContributi

metadataContribution

“contribution” appears in two LOM

categories defining a subclass for each

one;

.LearningObjectContribution  2.3. Lyfe

Cycle. Contribute

.MetadataContribution  3.2.

Metadata.Contribute

entityContribution, dateTimeContribution

lomAnnotation 8. Annotation

annotationDescription, annotationEntity,

date_that_this_annotation_was_created

lomClasification 9. Classification

clasificationPurpose clasificationTaxonPath,

taxonDescription

lomTaxonPath 9.2. Clasification. Taxon Path clasificationSource, clasificationTaxon

lomTaxon 9.2.2. Classification.Taxon Path.Taxon taxonEntry, taxonId

lomIdentifier

1.1. General.Identifier

3.1. Metadata.Identifier

7.3. Relation.Identifier

inCatalog, isEntry

lomRelation 7. Relation createdSource, relationKind

lomThecnicalRequiremen

4. Technical

singleTechnicalRequirem

ent

4.4. Technical. Requirement hasType, maxVersion, minVersion, requirementName

Vcard hasEmail, hasName, hasOrganization

learningObject

Learning object: This is the main class of

the ontology.

All the categories describe some property

of the learning object.

It has a lot of properties, one for each category and

subcategory of the object. This table only enumerates

some of them together with the corresponding IEEE LOM

category:

.hasAgregationLevel – 1.8. General. Aggregation Level

.hasCopyRightRestrictions – 6.2. Right. Right and

other restrictions

. hasCost – 6.1. Right.Cost

. hasDifficulty – 5.8. Educational. Difficulty.

. hasDuration – 4.7. Technical.Duration

.hasInstallationRemark – 4.5 Technical. Installation

remarks

.hasInteractivityLevel – 5.3. Educational. Interactivity

level

.hasInteractivityType – 5.1. Educational. Interactivity type

.hasOtherPlatformsRequirement – 4.6. Technical.

Other platform requirements

. hasSemanticDensity – 5.4. Educational. Semantic

Density

. hasSize – 4.2. Technical.Size

. hasStatus – 2.2. Life Cicle. Status

. hasStructure – 1.7. General. Structure

. hasTitle – 1.2. General. Title

.hasTypicalLearningTime – 5.9. Educational. Typical

learning time

. hasVersion – 2.1. Life cicle.Version

.hasMetadataLanguage – 3.4. Metadata. Language

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

320