Linked Data Strategy to Achieve Interoperability in Higher

Education

Guillermo García Juanes, Alioth Rodríguez Barrios, José Luis Roda García, Laura Gutiérrez Medina,

Rita Díaz Adán and Pedro González Yanes

School of Computer Science, University of La Laguna, San Cristóbal de La Laguna, Spain

Keywords: Linked Data, Open Data, Interoperability, High Education.

Abstract: An important challenge in centres of higher education is the use of Linked Data strategy to connect

currently existing multiple information systems. These information systems are usually independent from

one another, and the ability to obtain information by connecting different sources of data involves, in most

cases, unacceptable costs and effort. In this work, we have developed a platform based on Linked Data that

permits the interoperability of different sources of data, both internal as well as external. This

interoperability is achieved by 1) the use of higher education ontologies, and 2) the use of a process that

begins with the analysis of the data sources to be connected, followed by mapping of the closest ontologies,

and ends with the generation and publication of data in valid formats for Linked Data. The final product

permits stakeholders inside and outside the university to be able to make queries of two or more datasets in

different information systems at the same time.

1 INTRODUCTION

The term Open Data concerns offering to society the

data collected by public institutions, which, when

handled by third-parties, can be of great value for

the development of applications, reports, etc. The

principal objective of the Open Data strategy is to

offer transparency, participation and collaboration in

the publishing of information, in standard formats,

open and interoperable, facilitating its access and

permitting its re-use (Office, 2012). This is nothing

more than data that belong to public administration

being used, by individuals or companies, with or

without commercial ends, provided that use does not

constitute public administration activity. There are

many institutions that currently publish open data in

different formats (Fundación CTIC, 2013) (Bauer

and Kaltenböck, 2012).

The objective of Linked Data is to give meaning

to connections that are found in different datasets so

that machines can obtain more relevant information

making use of techniques from the Semantic Web

(Berners-Lee, 2006).

The relationship between Open Data and Linked

Data was proposed by Tim Berners-Lee who

suggested a way of measuring the degree of quality

of data published from an Open Data portal, where,

if those data were to be published using Linked Data

principles, the highest degree that a portal may

achieve would be obtained (Berners-Lee, 2006).

The data interoperatibility is achieved through

the followig of the Linked Data principles. These

vocabularies must cover a wide range of concepts

related to the university´s system. From the

institution, the departments, the teachers, including

the courses, programmes and study material, credits,

theory classes, practical classes and laboratory

classes, etc., a vocabulary must cover all these

aspects to be able to be linked and make full use of

the Linked Data strategy.

In this work we present a prototype that was

developed at the University of La Laguna (ULL),

where various different groups played a part: the

Planning and Analysis Office, the ULL Information

Technology Service, and the Taro Research Group.

The project concerns the demonstration of how

Linked Open Data offers a range of benefits for the

procurement of information that are not achieved

through other conventional methods. The higher

education system comprises multiple information

systems, some of which are quite complex. This

García Juanes G., Rodríguez Barrios A., Roda García J., Gutiérrez Medina L., Díaz Adán R. and González Yanes P..

Linked Data Strategy to Achieve Interoperability in Higher Education.

DOI: 10.5220/0004853500570064

In Proceedings of the 10th International Conference on Web Information Systems and Technologies (WEBIST-2014), pages 57-64

ISBN: 978-989-758-024-6

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

paper is, therefore, concerned with the application of

Linked Data strategy, methods and techniques to

some of these university systems.

2 RELATED WORK

The development of a platform such as the one

intended requires a prior state-of-the-art study, with

particular emphasis in the area of ontologies that

may be re-used with those that are going to model

concepts within the scope of higher education.

In the search for ontologies closer to our

problem, we made use of semantic search engines

such as Watson

or Swoogle

apart from search

engines for key words or known repositories like

Linked Open Vocabularies (LOV)

There is a great set of ontologies related to the

field of education, but we limited our search for

ontologies to those that could be adapted as far as

possible to our case. They were also evaluated for

their quality based on whether they were structured,

well documented and in current use by other

organizations.

From this analysis, candidate ontologies were

obtained for re-use in our system. The most relevant

ones, related to universities, were Academic

Institution Internal Structure Ontology (AIISO)

Teaching Core Vocabulary Specification (TEACH)

and The Bowlogna Ontology

. AIISO describes the

organizational structure of the university very

simply, showing the hierarchy and relationships

between different agencies. In that regard, it

provides classes for modelling: teaching staff,

subjects, departments, centres, etc. The TEACH and

Bowlogna ontologies describe the part more related

to teaching, that is, they try to represent the

relationship between a program, the subjects that

comprise it, the teaching workload and

responsibilities of teachers for each of the subjects.

Bowlogna is a more special case, representing the

organization of the university following the Bologna

Plan currently being implemented by European

universities (Demartini et al., 2013).

All things considered, a university is still an

organization, aside from the academic element,

which can be modelled with organizations

   

http://watson.kmi.open.ac.uk/WatsonWUI/

http://swoogle.umbc.edu/

http://lov.okfn.org/dataset/lov/

http://vocab.org/aiiso/schema

http://linkedscience.org/teach/ns/#

http://diuf.unifr.ch/main/xi/bowlogna

ontologies like W3C’s The Organization Ontology

or Buildings and Rooms Vocabulary

. Finally, we

intend to define the situation for premises and staff,

and CTIC’s Vocabulary of Localizations

or vCard

Ontology

can be used for that, providing a series of

classes with which to model addresses,

municipalities, provinces, etc. These ontologies are

very important as in many cases they are standard

and widely used, as their use is not limited to a

specific field.

Another important action was to take examples

of the strategies established by other universities as a

step towards the publication of Linked Data. They

usually follow a basic scheme, re-using as far as

possible existing ontologies, and if these do not

cover all of the necessary concepts, the ontology

itself is created or extended from an existing one.

This can be seen in the Open Data sections of the

University of Oxford

, the University of

Southampton

and the Open University

, that

usually have an Open Data portal from which you

can consult information modelled in different forms

(navigation, queries, etc.). These portals are usually

based on RDF software, Virtuoso being the most

popular, although there are other tools such as Pubby

, Fuseki

or D2R

The general situation is that when choosing

ontologies that adapt themselves more to

universities, there is consensus for the use of several

of those named above, but none ever achieves a

complete representation of the information, which is

why it is necessary to create another to be able to

link it with the other ontologies.

To achieve the publication of data, the ontologies

must be published in such a way that they are

dereferenceable and well documented, permitting

the rest of the world to re-use them, thus achieving

interconnections and future inferences of

information (W3C, 2008) (Heath and Bizer, 2011).

In our case, we do not enter into the creation of

ontologies in depth, as that was not one of the

principal objectives that we wished to demonstrate,

but rather interoperability between systems that

http://www.w3.org/TR/vocab-org/

http://vocab.deri.ie/rooms

http://purl.org/ctic/infraestructuras/localizacion

http://www.w3.org/TR/vcard-rdf/

https://data.ox.ac.uk/

http://data.southampton.ac.uk/

http://data.open.ac.uk/

http://wifo5-03.informatik.uni-mannheim.de/pubby/

http://jena.apache.org/documentation/serving_data/

http://d2rq.org/

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

implement Linked Data as a procedure prior to

publication of data, bringing it into a real context.

3 MOTIVATION AND GOALS

The University of La Laguna is a large-scale public

higher education institution. It currently involves

more than 26,000 people including students,

teachers and administration staff, distributed

between 25 centres, 60 departments and other areas.

There are also close links with private entities

(companies, foundations and institutes) and public

institutions such as the local island administration

Cabildo de Tenerife, City Councils and the

Government of the Canary Islands.

The organizational structure is decentralized, and

many functions are delegated to each of the

departments, centres and services. This fact has a

special relevance to this work, as we have had to

study the functions of the principal organizational

units in depth. Each unit works almost

independently, each one takes responsibility for

handling administrative processes that have been

delegated to them, collecting and maintaining the

information necessary to operate (accounting,

academic administration, libraries, ITC centres,

etc.).

With respect to this work, there are many

information systems that offer support to the

university as a whole. Financial Management and

Academic Management are among the main ones

that are found. The first system takes care of the

management of the administration staff and the

teaching staff; while the second is concerned with

the enrolment processes and everything related to

teaching. Both aggregate a large quantity of data and

are more or less controlled forming a fairly

homogenous architecture. They use the same group

of software development technologies, the same

database management system, and, most relevant,

they are under the responsibility of the same IT

department. Although there is a certain homogeneity

between these systems, extracting information

across both systems continues to be a complex and

costly task.

Apart from larger, older systems, there are

smaller, independent systems, of great value to the

institution. These have appeared over time according

to the needs of services or departments. Examples of

these systems are: the research service, the

directories of the institution´s staff, quality control,

diaries and events, etc. These are usually controlled

by different areas and each one can have different

software.

It is at this point that this work begins to have

meaning. There are many cases where management,

statistical or other similar information is requested

from other institutions within the university itself.

Most of these systems work independently from one

another, and when it is necessary to consult

information from two or more sources, it is

necessary to establish connections between the

different systems. Due to the complexity and

internal structure of each system, staff have to make

a particularly strenuous effort to obtain the data.

The main motivation to work in this environment

and to offer a practical solution to these problems

based on Linked Data is, on the one hand, the

existence of the real problem of access to different

sources of data, and on the other, that the

university´s staff is quite able to accept new

proposals and finally, the existence of a real and

concrete problem with which to apply Linked Data

methods and techniques (for the re-use of

ontologies, generation of RDF links, publication of

data and consumption of published data).

The key challenge is to offer the university a

solution for offering information obtained from a

variety of sources, without modifying the current

systems, as many are legacy systems and are so

assimilated within the institution that a change to

them could cause chaos. This is the solution that we

present below.

4 LINKED DATA PROCESS

The process of linking data from the different

information systems has followed an iterative and

incremental methodology (Suárez de Figueroa

Baonza, 2010). This process has allowed us to

obtain valuable results from the first iterations and

refine them continuously. As a consequence, the

Linked Data process is provided with the necessary

flexibility to tackle the changes related to the

requirements in any phase of the process.

The working methodology comprises six

differentiated phases inspired by methodologies

commonly used for the publication of Linked Data

(Poveda-Villalón, 2012); (Corcho et al., 2013);

(Fernández-López, 1997); (Atemezing et al., 2012)

and adapted to the needs of the particular working

environment. These phases are divided between:

specification of data to be published, data modelling,

generation of data in RDF, publication, linking and

exploitation.

Given the peculiarities of linking data in our

LinkedDataStrategytoAchieveInteroperabilityinHigherEducation

university context, as opposed to that proposed for

different methodologies for the publication of data

(Poveda-Villalón, 2012), the publication phase is

undertaken at an earlier stage to the linking of data.

4.1 Specification

The first task to be dealt with is the specification of

the data to be published. ULL has a large amount of

data related to administrative, academic or financial

activities. With the aim of demonstrating the utility

of linking data in the University under the paradigm

of Linked Data, and in view of the difficulty of

linking all the data in the institution owing to the

large volume, two clearly defined organizational

units were chosen: the organizational area and the

academic area.

The data available in each of these areas were

not related with one another, although they made

reference to entities that could be easily linked.

The data of the University´s organizational area

describe the hierarchy of the institution, the structure

of the teaching staff and the distribution of the

courses and the teaching areas.

Figure 1 presents

the organizational domain model:

Figure 1: Organizational domain model.

The academic area contains information relating to

the courses, which have been adapted to the Bologna

Plan. The data from the academic area, Figure 2,

contains information relating to the study plan of the

courses, provided by the Spanish Government´s

Ministry of Education, Culture and Sport, plus

information relating to the teaching staff teaching

those courses.

Figure 2: Academic domain model.

4.2 Modelling

Once the data to be linked had been determined, the

next step was to analyse them in order to be able to

begin modelling it with a set of ontologies.

Due to the peculiarities of the starting dataset, we

opted for the development of an ontologies network,

which had been re-utilized and created with

ontologies with different characteristics. This

ontologies network re-utilizes ontologies from the

academic environment such as AIISO and TEACH,

and more general and frequently used ontologies

such as FOAF

, SKOS

and others including LOC,

ORG and ROOMS that model localization, and

organization and premises, respectively. To

complete the network, we created three new

ontologies intended to represent the semantics which

earlier ontologies could not cover. One of the

ontologies was centred on academic information,

and the other on organizational information, and the

final one modelled general aspects of the institution.

The development of the network of ontologies was

carried out using NeOn Toolkit

, an open source

tool based on the Eclipse platform. It provides a set

of plug-ins that cover many of the needs arising

during this cycle of ontology development, such as

http://xmlns.com/foaf/spec/

http://www.w3.org/2008/05/skos

http://neon-toolkit.org/

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

the generation of documentation, modularization, or

the evaluation of ontologies (Zemmouchi-Ghomari

& Ghomari, 2013). Figure 3 shows the network

ontology obteined.

Figure 3: Modelled ontology network.

As part of the modelling phase, the anatomy of the

URI is defined under the guidelines of the so called

Cool URIs (W3C, 2008). As a result, the URIs of the

resources follow the following pattern:

http://datos.ull.es/resource/{resource

type}/{resource}

Code 1: URI structure.

For example, the “IT Engineering” course is

identified by the URI:

http://datos.ull.es/resource/programme/

IT_Engineering

Code 2: Example of an URI. IT Engineering identifier.

4.3 Generation

Once the ontology network has been defined, the

next step is to generate the data in standard RDF

format.

The D2RQ platform was used for that process,

which allowed for data stored on relational databases

to be consulted as if they were RDF graphs, making

use of SPARQL language. Using a mapping

language, in which the transformations to be made

were specified following the defined ontology, this

tool makes it possible to obtain data in RDF format

by directly consulting a relational database.

As a consequence of using a tool with these

characteristics, the results of this phase did not

consist of a set of data in RDF, but in a file with the

definition of the correspondences or mappings

between the different fields in the organizational and

academic area databases, and the elements of the

ontology network previously defined (see Figure 3).

An extract of the mapping file obtained during

this phase can be seen below:

map:institutions a d2rq:ClassMap;

d2rq:dataStorage map:database;

d2rq:uriPattern

“institutions/@@UNIVERSIDAD.NOMBRE|urli

fy@@”;

d2rq:class ullorg:Universidad;

Code 3: Mapping an institution type resource in D2R.

4.4 Publication

The main objective of this phase is to obtain two

access points for queries to the RDF format data of

the organizational and academic areas.

Thanks to the implementation of a Pubby based

interface, through the use of D2R, a front-end for

SPARQL endpoints, access can be given to data

stored in the institution´s relational databases

through an HTML interface which displays them in

RDF and at a SPARQL endpoint. This permits us to

navigate through the information as well as to make

specific queries.

4.5 Interlinking

Throughout this phase, we intended to achieve the

objective of the availability of a five star datasets,

following Tim Berners-Lee´s recognised

classification, that is, a set of data linked with other

sets of data (Heath and Bizer, 2012).

In our case, the linking process is divided into

two phases, one of which establishes a data link

internally, and another that links external data

sources.

The internal linking, between our two access

points, was achieved thanks to what was already

known about how the URIs were going to be

formed, as the structure is well defined and the

identifiers of each of the resources in most cases are

stipulated beforehand, and are common to the whole

organization. Due to this, it is not necessary to use

tools to discover candidates. Therefore is only

LinkedDataStrategytoAchieveInteroperabilityinHigherEducation

necessary that each system can be able to obtain the

identifiers from the other system to create the URIs.

In order to do this, the systems retrieve them from

the other source using a SQL script. This process is

only needed when new resource appears and not

each time that a query is done.

If this information was not held, it would be

necessary to make public the same resources in both

domains (to duplicate the occurrences) and then to

use linking tools, with owl:sameAs properties, to

indicate that they are the same resource.

The external linking is possible when external

data sources exist. For that, we undertook an

analysis, and due to the scarcity of reliable sources,

we limited ourselves to linking with dbpedia.org and

linkeddata.es.

This link entails a process in which candidate

links have to be found, and we intended to automate

it as far as possible, using a Silk tool

and Link

Specification Language (LSL) configurations. Once

the tool was implemented, results that corresponded

with the external resources that could be linked to

our data were generated. These links were added to

the database LINKED table in the following form:

http://opendata.ull.es/resource/{resour

ceid} owl:sameAs http://{external

resource}

Code 4: Structure of sameAs statement.

This table is like a repository or cache where all the

links that are found are stored and managed,

containing: the URI of the internal resource, the URI

of the external resource and the property that links

them. In this way we are able to make links not only

using owl:sameAs but also other useful properties,

as well as management fields for configuration

issues. This table also allow us to determine if a link

is valid, if has been manually blocked, the date that

the link was discovered, etc.

Once the links are added, the information is

republished automatically with the external links.

This is thanks to the added dynamic property in the

D2R map that takes care of reviewing the linked

table and showing the properties that exist there:

d2rq:belongsToClassMap map:CLASSTOLINK;

d2rq:dynamicProperty

"@@linked.propiedad@@";

d2rq:uriColumn "linked.objeto"; .

Code 5: Mapping sameAs statement in D2R.



http://wifo5-03.informatik.uni-mannheim.de/bizer/silk/

4.6 Exploitation

When addressing this phase, we started from the two

SPARQL access points available for the consultation

of data from the University´s organizational and

academic areas.

The objective of this phase was the availability

of a single point of access to the platform that would

permit federated queries on the two groups of data.

To achieve that objective, a Fuseki server was

deployed, allowing serve data in RDF format using

HTTP. This service offers data to users through a

RESTful API and an enriched SPARQL endpoint,

from which so called SPARQL++ queries may be

made (Polleres et al., 2007). With this service,

complex queries can be made using proprietary

databases or external endpoints. Previously, these

types of queries were only possible using scripts and

by the acquisition of data from sources outside the

organization.

Here is an example of a complex query making

use of two data sources:

PREFIX coreull: < … >

PREFIX orgull: < … >

PREFIX acaull: < … >

SELECT ?NameDept (COUNT(?planning) as

?numberPlanning) WHERE {

SERVICE <http://orgull/sparql> {

?teacher orgull:miembro ?dept .

?dept coreull:nombre ?NameDept .

}

SERVICE <http://acadull/sparql> {

?planning acaull:tieneProfesor

?teacher .

}

} GROUP BY ?NameDept

ORDER BY ?NameDept .

Code 6: Query example: “How many subjects does each

department teach?”

5 FINAL PRODUCT

The project finally developed is presented in Figure

4. We reproduced the complete system on a

development server simulating the real system used

at ULL. As it can be seen in the figure, we have two

independent information systems: Academic and

Organizational Management. Both systems reside in

different databases supported by MySQL. An

architecture based on D2R was developed above

each of the two current systems for the mapping of

the relational database to RDF files. Each system

physically resides in a different server and has,

WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies

therefore, a separate portal for each environment.

Each D2RQ server had an SPARQL endpoint

added and a web interface (Pubby-type) where

corresponding information could be obtained.

A supervised process to discover and update

external links was added using the Silk tool. More

over, Fuseki service was also added as a junction for

the whole system, enabling SPARQL++ queries.

Figure 4: Architecture diagram.

Finally, it was possible to obtain an open data

publication portal based on web semantics, meeting

Linked Data requirements. Users will be able to use

it to consult University information as well for

making queries at several SPARQL points (so called

SPARQL Endpoints) that may be linked to the

published data. An example of this would be making

a query such as “Distribution of students by

province and their geo-localization to be shown on a

map”, represented in SPARQL as:

PREFIX coreull:

<http://orgull/ontologies/core#>

PREFIX orgull:

<http://orgull/ontologies/organization#

PREFIX acaull:

<http://orgull/ontologies/academic#>

PREFIX georss:

<http://www.georss.org/georss/>

SELECT ?Region (COUNT(?student) as

?numberStudents) ?geoloc

WHERE {

SERVICE < http://acadull/sparql >

{

?student a acaull:Alumno;

coreull:provincia ?reg .

}

SERVICE < http://orgull/sparql > {

?reg coreull:nombre ?Region ;

owl:sameAs ?linked .

}

SERVICE <

http://dbpedia.org/sparql > {

?linked georss:point ?geoloc .

}

} GROUP BY ?Region .

Code 7: Query with external data sources. Recount of

students and their geolocalization.

The final platform enables users with simple

SPARQL queries to make cross-queries of data from

different sources that were previously inaccessible

or of very complex interoperability. The end users,

both inside and outside the University, have

available a tool based on Linked Data to obtain the

information they need. If case of need, it would be

possible to add a user interface to make its use more

user-friendly.

6 CONCLUSIONS

In this work, we have developed a genuine platform

based on the Linked Data strategy in which

universities may publish data and link with internal

and external sources. The interlinking of the

different university data sources will permit greater

interoperability and can achieve new knowledge

from this relationship.

To deploy our prototype in the university

production servers, it is necessary to consider two

important aspects: the integration costs of our

platform with the university systems and also the

university staff training in linked data. We firmly

believe that both aspects can be addressed as we

have the necessary technical expertise.

Whit this project, the institution has a tool to

make complex queries, which hitherto existing

systems could only handle with great effort.

Moreover, the same tool could be offered to other

local, island, and regional institutions in such a way

that they could make queries directly, rather than

through a service request to university personnel.

The ontologies most related to higher education

have been revised, and improvements have been

proposed to cover concepts not covered by existing

ontologies. The network of ontologies used in this

work covers ten ontologies, of which only three have

been newly created.

Finally, as a summary of the work undertaken,

we would stress that effective analysis of the initial

data produces a lower number of errors in the re-use

and definition of ontologies. We would also indicate

that the process of creation of ontologies was based

on the basic terms needed to demonstrate the

viability of the project. We will continue to research

several concepts that are beginning to be requested,

in more depth, once the first version of the platform

has been validated. With respect to the technological

platform, a viability study will have to be undertaken

to compare the effort required to incorporate triple

store systems like Virtuoso, etc.

LinkedDataStrategytoAchieveInteroperabilityinHigherEducation

We should not forget that any new system for the

organization would also implicate the commitment

of human resources and materials for its

development and future maintenance.

ACKNOWLEDGEMENTS

We would like to thanks to Andrés Palenzuela from

the Planning and Analysis Office of University of La

Laguna in the Canary Islands. His interest and

support has made this Project viable.

REFERENCIAS

Atemezing, G., Corcho, O., Garijo, D., Mora, J., Poveda-

Villalón, M., Rozas, P., Villazón-Terrazas, B. (2012).

Transforming meteorological data into linked data. En

Semantic Web. IOS Press.

Bauer, F., & Kaltenböck, M. (2012). Linked Open Data:

The Essentials A Quick Start Guide for Decision

Makers.

Berners-Lee, T. (2006). Linked Data. Recuperado el 11 de

10 de 2013, de http://www.w3.org/DesignIssues/

LinkedData.html

Corcho, O., Fernández-Lopez, M., & Gómez-Perez, A.

(2003). Methodologies, tools and languages for

building. Where is their meeting point? En Data &

Knowledge Engineering (págs. 41–64). Elsevier.

Demartini, G., Enchev, I., Gapany, J., & Cudré-Mauroux,

P. (2013). The Bowlogna Ontology: Fostering Open

Curricula and Agile Knowledge Bases for Europe’s

Higher Education Landscape. Semantic Web –

Interoperability, Usability, Applicability, 4(1), 115.

Fernández-López, M. y.-P. (1997). METHONTOLOGY:

From Ontological Art Towards Ontological

Engineering. AAAI-97 Spring Symposium Series,

(págs. 24-26). Stanford University, EEUU.

Fundación CTIC. (2013). Public Dataset Catalogs

Faceted Browser. Recuperado el 11 de 10 de 2013, de

http://datos.fundacionctic.org/sandbox/catalog/faceted/

Heath, T., & Bizer, C. (2011). Linked Data: Evolving the

Web into a Global Data Space (First ed.). Morgan &

Claypool.

Office, C. (2012). Open Data White Paper: Unleashing

the Potential. TSO (The Stationery Office).

Polleres, A., Scharffe, F., & Schindlauer, R. (2007).

SPARQL++ for Mapping Between RDF Vocabularies.

En On the Move to Meaningful Internet Systems 2007:

CoopIS, DOA, ODBASE, GADA, and IS (págs. 878-

896). Springer Berlin Heidelberg.

Poveda-Villalón, M. (2012). A Reuse-Based Lightweight

Method for Developing Linked Data Ontologies and

Vocabularies. En The Semantic Web: Research and

Applications (págs. 833-837). Springer Berlin

Heidelberg.

Suárez de Figueroa Baonza, M. d. (2010). NeOn

Methodology for Building Ontology. M.C. Doctoral

Thesis. Universidad Politécnica de Madrid.

W3C. (2008). Best Practice Recipes for Publishing RDF

Vocabularies. Recuperado el 11 de 10 de 2013, de

http://www.w3.org/TR/swbp-vocab-pub/

W3C. (2008). Cool URIs for the Semantic Web.

Recuperado el 11 de 10 de 2013, de

http://www.w3.org/TR/cooluris/

Zemmouchi-Ghomari, L., & Ghomari, A. R. (2013).

Process of Building Reference Ontology for Higher

Education. World Congress on Engineering 2013, III,

1595-1600.



WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies