ArticlePDF Available

Linked Data in Education: A Survey and a Synthesis of Actual Research and Future Challenges

December 2017
IEEE Transactions on Learning Technologies 11(3):1-1

December 2017
11(3):1-1

DOI:10.1109/TLT.2017.2787659

Authors:

Crystiam Pereira e Silva

Sean Wolfgand Matsui Siqueira

Universidade Federal do Estado do Rio de Janeiro (UNIRIO)

Bernardo Pereira Nunes

Australian National University

Stefan Dietze

GESIS - Leibniz-Institute for the Social Sciences

Linked Data principles and technologies are being investigated in various areas. In the Educational context, many studies are using Linked Data trying to solve problems of interoperability of educational data and resources, enriching educational content, and personalizing and recommending educational content and practices. This article presents a systematic mapping of proposals which have been adopting Linked Data for supporting education, and, based on analysis of these proposals, we discuss the tools, vocabularies, and datasets being used, providing a research landscape of the area. We also present challenges and trends which can foster future research in this area. IEEE

NUMBER OF PAPERS PER YEAR

…

DISTRIBUTION OF PAPERS BY RESEARCH DATABASE

…

NUMBER OF STUDIES ANALYZED AT EACH STAGE

…

SOME INSTITUTIONS THAT ARE RESEARCHING AND APPLYING LINKED DATA TECHNOLOGIES IN THE EDUCATIONAL CONTEXT.

…

Figures - uploaded by Bernardo Pereira Nunes

Content may be subject to copyright.

Content uploaded by Bernardo Pereira Nunes

Content may be subject to copyright.

information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TLT.2017.2787659,

IEEE Transactions on Learning Technologies

IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID 1

Linked data in Education: a survey and a

synthesis of actual research and future

challenges

Crystiam Kelle Pereira, Sean Wolfgand Matsui Siqueira, Bernardo Pereira Nunes, Stefan Dietze

Abstract—Linked Data principles and technologies are being investigated in various areas. In the Educational context, many

studies are using Linked Data trying to solve problems of interoperability of educational data and resources, enriching

educational content, and personalizing and recommending educational content and practices. This article presents a systematic

mapping of proposals which have been adopting Linked Data for supporting education, and, based on analysis of these

proposals, we discuss the tools, vocabularies, and datasets being used, providing a research landscape of the area. We also

present challenges and trends which can foster future research in this area.

Index Terms— Semantic Web, Semantic technology, Educational technology

——————————  ——————————

1 INTRODUCTION

inked Data (LD) can be summarized as the use of the

Web to create connections between data which may

be originally stored in various databases maintained by

different organizations and distributed across different

geographic locations [1]. One of its primary objectives is

to extend the Web of Documents, where HTML documents

are interconnected through hyperlinks, to a Web of Data

[1], where data may be connected directly, following the

LD principles [2]. The LD principles can be summarized

as follows:

1. use URIs as names for things;

2. use HTTP URIs so that people can search for the-

se names;

3. provide useful information when someone

searches for a URI, using recommended stand-

ards (RDF - Resource Description Framework [3],

SPARQL – Simple Protocol and RDF Query Lan-

guage);

4. add links to other URIs so that additional infor-

mation can be discovered.

A few years after these first concepts were presented,

the “5 stars of Linked Data” were defined by Tim Bern-

ers-Lee [2], classifying solutions so that the more stars a

solution has, the more powerful and easy to use it is.

Although this area presents many limitations and chal-

lenges, the use of LD has brought numerous benefits,

including transparency, reusability, knowledge discov-

ery, and interoperabilityfor various application areas.

For instance, LD has been used in various fields, in-

cluding the area of Education. Linked Data has potential

for use in Education due to the open and accessible na-

ture of the educational data and resources produced by

many institutions [4]. It can be an alternative for sharing

and reusing educational data and resources, providing

interoperability among repositories, enriching content,

exploring large datasets relevant to education perfor-

mance analysis, individualizing and personalizing learn-

ing, as well as other issues which are still challenges. The

ability to share and connect data has encouraged many

educational institutions to begin a movement to use

Linked Data principles and technologies in education.

Some notable projects in this regard are LinkedUp pro-

ject

, Linked Education Cloud

, mEducator

, Open Uni-

versity

, LAK Dataset

and LAK Data Challenge

Different initiatives seek to share data such as courses

offered by universities [5][6], statistical data [7], organiza-

tional data [8][9], educational resources such as videos,

presentations, lectures, books and games [10][11][12], as

well as the publication of tools to support educational

practices [13][14].

In addition to providing data in a structured way, data

integrating with other linked datasets and data reusing

for enrichment [15][16][17][18][19][20], recommendation

(and customization) of educational content available in

the Web of Data [21][22][23][24][25], and expansion of

search terms [26][27] are the main objectives of the work

analyzed. The increase in exposing educational data

through Linked Data technologies creates also

the opportunity for the development of applications in

the learning analytics (LA)/educational data mining

(EDM) field [28][29][30].

https://linkedup-project.eu/

http://data.linkededucation.org/linkedup/catalog/

http://www.meducator3.net/

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

https://solaresearch.org/initiatives/dataset/

http://meco.l3s.uni-hannover.de:9080/wp2/?page_id=18

xxxx-xxxx/0x/$xx.00 © 200x IEEE Published by the IEEE Computer Society

————————————————

 Crystiam Kelle Pereira and Sean Wolfgand Matsui Siqueira and Bernardo

Pereira Nunes are with the Department of Applied Informatics (DIA), Fed-

eral University of the State of Rio de Janeiro (UNIRIO), Av. Pasteur, 458,

Urca, 22290-240 Rio de Janeiro, RJ, Brazil. E-mail: {sean, crystiam.kelle,

bernardo.nunes}@uniriotec.br

 Bernardo Pereira Nunes is also with the Department of Informatics - PUC-

Rio, Rua Marquês de São Vicente, 455, Gávea, Rio de Janeiro, RJ, Brazil.

Email: bnunes@inf.puc-rio.br

 Stefan Dietze is a Research group leader at the L3S Research Center of the

Leibniz University Hanover, Germany. Email: dietze@l3s.de

information.

IEEE Transactions on Learning Technologies

2 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

The research and application of LD principles and

technologies are continually evolving. For this reason,, it

is important that current initiatives are mapped to indi-

cate opportunities, identify technologies, approaches and

tools available, report experiences and address challenges

and trends for future research.

To map out this research landscape, this article con-

ducts a systematic mapping of state of the art in LD in

Education. It collects the various ways in which LD is

being used in the context of Education, for instance, the

tools supporting the processes of data publishing and

reuse, as well as the relevant vocabularies and datasets. In

addition, the challenges currently faced in each stage of

the process are raised. Finally, to provide an overview of

the current developments in the area, this study can also

serve to guide future work and to suggest possible ave-

nues for further research.

The remainder of this paper is organized as follows.

Section 2 reviews the literature. Section 3 presents the

systematic review and mapping. Section 4 overviews the

use of LD in Education. Section 5 presents datasets and

vocabularies commonly used. Section 6 presents the chal-

lenges faced and, finally, Section 7 concludes the paper.

2 RELATED WORK

The application of LD in Education was discussed as

early as 2009 [31]. On that occasion, the authors reviewed

the status of this research area, examining the pioneering

initiatives, the main challenges, and the datasets, tools,

and prominent applications.

The application of semantic technologies in education-

al environments on the Web was also the subject of a

survey in 2011 [32]. The central axis of that investigation

was a discussion of ontologies and vocabularies geared

towards the area of Education.

LD in Education is also discussed by D’Aquin et al. [4] as

a potentially revolutionary Web technology for problems

in the area of distance learning and open education. The

authors argue that the use of LD can help with naviga-

tion, delivery, discovery and recommendation of educa-

tional resources, and cite possibilities for customizing

learning and establishing social connections. Likewise,

D’Aquin et al. [4] describe how organizations are using

LD technologies. They also provide a brief overview of

the data publishing process, as well as key tools and

technologies involved in this process.

Another relevant survey in LD in Education is presented

by Dietze et al. [33] where the authors state that the appli-

cation of LD principles in education offers excellent po-

tential for addressing problems involving interoperability

of educational data and resources. They also discuss vari-

ous scenarios that can benefit from the use of LD in Edu-

cation as well as present an overview of the data publica-

tion and interlinking process and the various tools and

technologies adopted so far along with the main chal-

lenges faced on the use of LD in Education.

In another survey conducted by Navarrete and Lujan-

Mora [34] the topic of LD in Education is presented under

the perspective of the enrichment of open educational

resources, prioritizing the description of technologies and

proposals related to this specific topic.

Although there are other studies[33][34][35] that also

reviewed the use of Linked Data in Education, this paper

aims not only to provide an overview of the works in the

area, but mainly to understand the objectives of using

Linked Data in Education. It also analyses outdated and

trends on Linked Data tools and vocabularies in the Edu-

cational context. This paper is a supplement of previous

works presenting a comprehensive survey with a total of

114 papers reviewed out of 1384 pre-selected papers in

the field (see selection criteria in Section 3). The analysis

of a larger volume of papers provides more evidence

with regards to which LD initiatives in Education are

using tools, vocabularies, and datasets. Additionally, the

present paper provides an overview of the current chal-

lenges faced in each stage of the process of applying LD

in Education.

3 SYSTEMATIC REVIEW AND SYSTEMATIC

MAPPING

The selection of papers reviewed in this study was per-

formed using a systematic mapping process. According to

Kitchenham and Charters [36], a systematic mapping

should allow guiding the focus of future systematic re-

views while also identifying areas for further primary

studies to be conducted. As stated by Kitchenham and

Charters [36], “the well-defined methodology makes it

less likely that the results of the literature are biased, alt-

hough it does not protect against publication bias in the

primary studies”.

Therefore, following strictly the recommendations

proposed by Kitchenham and Charters [36], the present

survey was conducted using the following steps: planning,

implementation, and reporting.

3.1. Planning

According to Biolchini et al. [37], in the planning phase,

the research objectives are listed, and the research ques-

tions are formulated, thereby generating a search string.

The present systematic mapping was designed to an-

swer two primary research questions:

 (RQ01) What are the objectives that Linked Data is

being used for in the context of Education?

 (RQ02) What are the Linked Data datasets used in

the context of Education?

Based on analysis of the studies performed in this area,

we also verified some secondary questions:

 (SQ01) What ontologies and vocabularies are being

used in the publication of Linked Data in the con-

text of Education?

 (SQ02) What tools are being employed to support

the process of using Linked Data in Education?

Although the initial objective was to answer questions

associated with the goals, datasets, ontolo-

gies/vocabularies, and tools related to Linked Data in the

context of Education, the review of 114 papers, as well as

the numerous research references, allowed us to amplify

the discussion, resulting in an up-to-date survey of the

information.

IEEE Transactions on Learning Technologies

area.

So, the search queries were formulated in three stages:

(i) creating search questions based on PICOC (Population,

Intervention, Comparison, Outcome, and Context) [36];

(ii) identifying synonyms for the keywords; and (iii) con-

structing a search string.

Based on the definition of the research question, it was

possible to extract the keywords (linked data, education)

and their synonyms (linked open data, open data, learning).

The combination of these terms resulted in the following

search string:

(“Linked data” OR “Open data” OR “Linked Open Da-

ta”) AND (“Education” OR “Learning”)

During the development of the protocol, inclusion cri-

teria were also defined (proposals using LD in the context

of Education), as well as criteria for exclusion of articles,

as follows: (1) derivative works; (2) works not available in

English, Portuguese or Spanish; (3) articles lacking online

access; (4) short papers (articles having less than 4 pages);

(5) studies having technical problems in cataloging or

retrieval; (6) descriptions of workshops and/or confer-

ences; (7) studies not in the field of learning technologies;

(8) studies not related to LD; (9) studies not covering the

use of datasets in education; and (10) studies on LD

where the data was not actually linked.

In order to verify the quality of the search string, con-

trol articles were considered [9][38][39][40][41]. The ex-

pectation was that the control articles should be retrieved

using the defined search string. All the control articles

were indeed retrieved. Some of them were discarded in

the selection steps because they were secondary papers.

The tools used to support the implementation of the

systematic mapping were: (1) StArt

(State of the Art

through Systematic Review): a tool responsible for sup-

porting the entire mapping process, from planning to

final reporting, as well as the stages of filtering and select-

ing articles; (2) Mendeley

: used to support reading and

marking the papers retrieved and managing the refer-

ences; (3) Excel: to support extraction and manipulation

of data, and generation of reports.

3.2. Implementation of the Systematic Mapping

Once the protocol was defined, the systematic mapping

was started by applying the search string to the following

six databases: ACM Digital Library

, IEEE Xplore Digital

Library

, Springer

, Scopus

, ScienceDirect

and Web of

Science

. Adjustments in the search string were needed

for some databases to follow their specific formats, with-

out, however, changing the terms.

1,957 papers were retrieved, of which 573 were dupli-

cates, thus leaving 1,384 articles to be analyzed. The dis-

tribution of articles retrieved from each database is shown

in Table 1.

http://lapes.dc.ufscar.br/tools/start_tool

https://www.mendeley.com/

http://dl.acm.org/advsearch.cfm

http://ieeexplore.ieee.org/search/advsearch.jsp

https://www.springer.com

https://www.scopus.com/

http://www.sciencedirect.com/science/search

https://webofknowledge.com/

In the first step of the article selection process, the titles

and abstracts of the papers were read, resulting in 1,157

being discarded, and 227 being selected for the subse-

quent step of the more detailed analysis. A summary of

the number of studies at each step is shown in Table 2.

In the second stage, a closer analysis of the papers was

performed, by reading the abstracts and the whole paper.

111 studies were rejected using the exclusion criteria al-

ready mentioned. Also, two duplicate studies were identi-

fied. Thus, 114 papers were selected for the data extrac-

tion step.

An overview of the studies by date shows (Figure 1)

that the topic started being researched in 2009, with sig-

nificant growth in the following years through 2012. The

same behavior can be observed in the study by Vega-

Gorgojo et al. [35]. The topic reached a new level of

growth in 2015. However, it is not possible to state

whether there is a growth trend since the search cut-off

point for studies in 2016 was April of that year.

4. OVERVIEW OF LINKED DATA IN EDUCATION

According to Dietze et al. [33], the adoption of LD princi-

ples can benefit the broad integration of educational data,

which involves two steps:

1. Educational Data Integration: this deals with the

integration of heterogeneous educational data,

through its exposure and interconnection with

other data sets.

2. Educational Services Integration: this seeks to re-

solve heterogeneity between standards adopted by

different APIs (Application Programming Inter-

faces) and repositories.

3. Linked Data has a potential for use in Education

due to the open and accessible nature of the educa-

tional data and resources produced by many uni-

versities [4]. While the content of open educational

resources is, by definition, accessible and reusable,

publishing its corresponding metadata as LD can

make the content of different repositories more

discoverable, accessible and connectable.

2009 2010 2011 2012 2013 2014 2015 2016

Number of papers per year

FIGURE 1 - NUMBER OF PAPERS PER YEAR

information.

IEEE Transactions on Learning Technologies

4 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

TABLE 1

DISTRIBUTION OF PAPERS BY RESEARCH DATABASE

Research database

Number of papers retrieved

ACM

160

IEEE

122

Springer

361

Scopus

906

ScienceDirect

Web of Science

344

TABLE 2

NUMBER OF STUDIES ANALYZED AT EACH STAGE

SELECTION

EXTRACTION

INITIAL NUMBER OF STUDIES

1,957

227

ACCEPTED

227

114

DISCARDED

1,157

111

DUPLICATES

573

Various initiatives for publishing data on the Web in

an open, structured and connected manner are notewor-

thy in the context of education:

1. LinkedEducation.org

is a project focused on shar-

ing resources and information related to LD in

education.

2. The University of Southampton

provides public

access to some of its administrative data to allow

the construction of tools leveraging this data, with

the primary objective being transparency of its ac-

tions.

3. The Open University in the UK

was the first edu-

cational organization to create a LD platform ex-

posing information from all its departments,

courses, multimedia and scientific publications

[42].

Table 3 shows the main institutions, ordered by the

number of publications retrieved and analyzed, that are

researching and applying linked data technologies in the

educational context.

In this research, we examined 114 proposals to identify

what are the objectives that Linked Data is being used for in the

context of Education?” (RQ01), that is, how the linked data

are contributing to solving problems in the educational

field.

The analysis of the proposals indicates that the use of

Linked Data in Education has three main objectives:

1. Availability of educational data as Linked Data;

2. Educational data and systems integration and in-

teroperability from different sources; and,

3. Consumption of Linked Data for different educa-

tional purposes.

A general discussion about the objectives is presented

with examples of Linked Data use. Furthermore, this

section also provides an overview about “What tools are

http://linkededucation.org/

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

http://dados.open.ac.uk

being employed to support the process of using Linked Data in

Education?”(SQ02).

TABLE 3

SOME INSTITUTIONS THAT ARE RESEARCHING AND APPLYING

LINKED DATA TECHNOLOGIES IN THE EDUCATIONAL CONTEXT.

INSTITUTION

COUNTRY

L3S Research Center

Germany

National Research Council of Italy Inst. for

Educational Technology

Italy

Departamento de Ciencias de la Computaci-

ón Universidad Técnica Particular de Loja,

UTPL Loja

Ecuador

Democritus University of Thrace/School of

Medicine, Alexandroupoli

Greece

Facultad de Informática Universidad Politéc-

nica de Madrid

Spain

Dept of Electronics and Information Systems

– Multimedia Lab Ghent University – IBBT

Belgium

Olayan School of Business American Univer-

sity of Beirut

Lebanon

Medical Informatics Laboratory, School of

Medicine, Aristotle University of Thessaloniki

Greece

Web Science Program School of Mathematics

Aristotle University of Thessalonik

Greece

The Open University

Dipartimento di Ingegneria Elettrica, Elettro-

nica e Informatica, Università di Catania

Italy

School of Medicine, Democritus University of

Thrace

Greece

Information Engineering Research Unit,

Computer Science Department, University of

Alcalá

Spain

School of Telecommunication Engineering,

University of Valladolid

Spain

Computer Science Department of the Univer-

sity of Alcalá

Spain

4.1 Availability of educational data as Linked Data

The area of Technology Enhanced Learning has long been

striving for wider reuse and sharing of resources, and the

use of Linked Data has been an alternative.

Note that more than 63% of the studies are focusing on

issues related to the sharing and publishing of education-

al resources as Linked Data. This “trend” is quite under-

standable when considering two important aspects of this

process. The first aspect is that the initiative of providing

information.

IEEE Transactions on Learning Technologies

open data, in a structured format, using a semantic repre-

sentation, is still very recent. Many of the studies sur-

veyed are still in the initial stages of the process of pub-

lishing and using (consuming) LD. The second aspect is

that many data repositories, from different domains, must

go through a series of standardization, analysis and ad-

justment processes to be made available using the LD

principles. The challenges involved in this data prepara-

tion process motivate various studies.

According to the literature mapping, the nature of the

data being published is diverse: data from courses offered

by universities [5][6], statistical data [7], organizational

data [8][9], educational resources such as videos, presen-

tations, lectures, books and games [10][11][12], as well as

the publication of tools to support educational practices

[13][14].

The publication of data can be made following 5-star

deployment scheme suggested by Tim Berners-Lee [2].

Some tools can support the 5-star Linked Data publishing

process, which are organized according to their function:

tools for converting and mapping data to RDF, tools for

storing and accessing LD, and tools for data interlinking.

A. Tools for converting and mapping data to RDF

Many tools have been developed and made available to

assist data publishers in the process of transforming and

publishing data following LD standards. The transition

from data represented in spreadsheets, e-mails, or even in

relational databases is not trivial and demands that the

publishers know vocabularies that can describe their data

as well as how to connect their data to other different

datasets.

The D2RQ

platform [43] is the most used in the stud-

ies analyzed. The D2RQ platform has some tools for

transforming a relational database to RDF. These include

the D2R Server for publishing relational databases on the

Semantic Web and D2RQ mapping tool for the creation of

mappings between a relational database and a specific

RDF schema. For example, Rajabi et al. [44] use the D2RQ

platform to expose some learning objects elements of

GLOBE repository aiming at evaluating the interlinking

results between GLOBE repository and several educa-

tional datasets on the Web of Data.

Like D2RQ, Triplify

e OpenRefine

are used to pub-

lish RDF and Linked Data from relational databases.

Triplify [17] is used in the MetaMorphosis+ environment

[45] for publishing, sharing and repurposing educational

content in medical education. The Open Refine is used by

Penteado [30] to extract and convert the data from the

data sources, reconcile with DBpedia, and generate RDF

triples. (Answer R4C2)

Although there are tools to support the data publica-

tion step, they often do not capture complex domain se-

mantics that is required by many educational applica-

tions. A lot of work is needed to create mappings manual-

ly, as discussed by Sahoo et al. [46]. In September 2012,

W3C published the R2RML recommendation [47], a

http://d2rq.org/

http://semanticweb.org/wiki/Triplify.html

http://openrefine.org/

standard language to describe mappings between a rela-

tional database and an equivalent RDF dataset, encourag-

ing RDB-to-RDF tool developers to comply with a stand-

ard mapping language.

Because of the difficulty in finding solutions focused

on the educational context, some proposals have devel-

oped customized tools. Pirrotta [7] developed a tool to

convert and publish statistical data about University stu-

dent activities. Metadata mapping processes were pro-

posed by Thangsupachai et al. [48] to convert database to

RDF to Linked Open Data in a particular case of mathe-

matics courses, using a variety of tools.

B. Tools for storing and accessing Linked Data

After the data have been converted, a next step is to store

them. A triplestore

is a type of database management

system that uses RDF as its data model. These tools often

provide access to their data via SPARQL endpoints to

support queries. The Sesame

and OpenLink Virtuoso

were the most mentioned tools in the studies analyzed.

Other triplestore systems mentioned were: Allegro-

Graph

, Mulgara

, ARC2

, Apache Jena Fuseki

and

BigOWLIM

Note that triplestores are systems aimed to store and

retrieve data through semantic queries. Therefore, none of

the papers analyzed mentioned specific educational

needs. A comparative analysis between triplestore tools

can be found in [49][50].

Despite the increase of tools for publishing and storing

Linked Data on the Web, their use is not a trivial task and

requires technical knowledge and understanding of Se-

mantic Web concepts and technologies. Most of the pa-

pers analyzed needed to use several different types of

tools or even create tools to support the data publishing

process. Alcantara et al. [51] propose the development of

embedded friendly interface in data provider systems for

supporting the transformation in different formats.

On the other hand, the data publication demands

knowledge in the domain of education to lead the crea-

tion of mappings that encompasses educational issues.

The combination of efforts from both areas (education

and technology) can enhance the quality of some specific

tools and approaches for educational data publication. A

listing of some other tools can be found in several studies

[31][32][35][52][31][33][35].

4.2 Educational data and systems integration and

interoperability

One of the major challenge in the Learning Technologies

is integration and interoperability of educational systems,

resources and data. This challenge is due to the variety of

systems being used around the world, and the need for

reuse of educational resources, and data.

In recent years, many initiatives have sought to create

http://www.dataversity.net/introduction-to-triplestores/

https://www.w3.org/2001/sw/wiki/Sesame

https://virtuoso.openlinksw.com/

http://franz.com/agraph/allegrograph/

http://www.mulgara.org/

https://github.com/semsol/arc2/wiki

http://jena.apache.org/documentation/fuseki2/index.html

[Bishop et al. 2011]

information.

IEEE Transactions on Learning Technologies

6 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

vocabularies and standards that allow the integration and

interoperability of educational data. The multiple results

of these efforts eventually created different patterns, us-

ing different models, technologies, and languages.

As explained by Zouaq et al. [53] Linked Data has the

potential to be a possible solution by allowing the map-

ping between various data models. Thus, semantically

similar models that are represented differently can still be

aligned using links to establish meaningful connections

between concepts from different models.

Some proposals specifically dealt with investigating

the possibility of system and data interoperability using

LD technologies [13][54][55][56][57], through the inter-

linking between educational data and resources across

different datasets [54][55][57], generation of links between

different datasets, definition of mapping between differ-

ent educational datasets [56], evaluation of the interlink-

ing of two datasets [44][58].

A. Tools for Data interlinking

Once data are made available, it is recommended to con-

nect them with other datasets to achieve the highest level

of maturity in data provision [2]. Some of the tools men-

tioned in the studies analyzed that can support the data

interlinking process are: Silk

, LIMES

, UCI [59] and

RDF-IA [60].

According to Lehmann et al. [61], SILK and LIMES are

the most popular frameworks among the LD publishers.

SILK is used by Konstantinidis et al. [62] to discover rela-

tionships between the metadata of the educational re-

sources and other data sources on the Web and to create

RDF links to link them. LIMES is used by Rajabi et al. [44]

to interlink the GLOBE repository to 20 datasets in the

Linked Open Data (LOD) cloud. When a new dataset is

published on the Web of Data, it is important, as recom-

mended by [2], to identify other possible datasets to inter-

link with. However, identifying these datasets is not a

trivial task. For instance, Leme et al. [63] and Wölger et al.

[64] suggest candidate datasets recommendation tech-

niques based on the vocabularies, classes, and properties

commonly used. A comparative analysis of the interlink-

ing tools is provided by Wölger et al. [65] and Rajabi et

al.[66]. Techniques and common issues for the data inter-

linking process can be found in [20][52][58][67][68][69].

4.3 Consumption of Linked Data in Education

Once the data are made available, there are significant

opportunities for the development of innovative solu-

tions, either through analysis of the current educational

context and decision making more conscious and directed

to the revealed problems or by the development of appli-

cations that explore the data and resources available to

improve the quality of education.

Some of the papers analyzed in the systematic review

consume Linked Data from different datasets for different

purposes, such enrichment, recommendation and reuse.

Some examples are enrichment of books through associa-

tion with other data [15][16][17][18][19][20], the use of

http://silkframework.org/

http://aksw.org/Projects/LIMES.html

collection of a museum or library, made available as LD

to support the explanation of a particular subject in the

classroom, or preparing questionnaires through datasets

of questions and tasks [70][71], recommendation (and

customization) of educational content available in the

Web of Data [21][22][23][24][25], and expansion of search

terms [26][27].

To semantically enrich an educational resource, sys-

tems often extract structured information from unstruc-

tured data and link to external knowledge bases in the

Linked Open Data cloud (LOD), such as DBpedia. This

annotation process can benefit systems, enhance infor-

mation retrieval and improve interoperability. Infor-

mation retrieval is improved by the ability to perform

searches which exploit the ontology to make inferences

about data from heterogeneous resources; furthermore,

annotations based on a common ontology can provide a

common framework for the integration of information

from heterogeneous sources [72].

A. Linked Data for Learning Analytics

The increase in exposing educational data through Linked

Data technologies creates the opportunity for retrieving

information and knowledge about educational resources,

learning processes, data interlinking between different

sources and analysis of educational content. These many

possibilities may contribute to the development of appli-

cations in the learning analytics (LA)/educational data

mining (EDM) field.

Zouaq et al. [53] discuss some opportunities and chal-

lenges associated with the use of Linked Data in educa-

tional data integration and analysis. They argue that the

use of LD is particularly relevant for the LA field because

it enables the collection and management of learner and

content data from a variety of sources (applications and

services) used in informal and lifelong learning.

The Learning Analytics and Knowledge (LAK) da-

taset

[28] exposes a collection (over five years) of biblio-

graphic resources about Learning Analytics and Educa-

tional Data Mining. The dataset reuses vocabularies and

provides links to schemas and entity coreferences in re-

lated datasets. Penteado [30] and Maturana et al. [29]

created applications that allow users to perform data

analysis and to explore data about researchers, confer-

ences, and publications available in the LAK dataset.

The application of LD in learning analytics (LA) is the

subject of several proposals. Reflection and prediction of

trends in Personal Learning Environment (PLE) is a pri-

mary focus of [23]. They are using semantic context mod-

eling and creation of Linked Data from logs. Mey-

mandpour and Davis [73] developed a metric to create a

rank of universities based on educational Linked Data.

The proposed metric uses the quality and value of some

semantic properties to identify the relative position of

universities worldwide. Penteado [30] uses both educa-

tion and economic indicators for analysis of school per-

formance of Brazilian schools. The datasets containing the

indicators were made available as Linked Data and en-

riched by indicators information extracted from DBpedia,

http://lak.linkededucation.org/

information.

IEEE Transactions on Learning Technologies

allowing a comparative analysis.

B. Tools for Consumption of Linked Data

DBpedia offers some tools to support the document en-

richment process. DBpedia Spotlight

adds markups

with semantic information surrounding atomic elements

(entities). These entities are the ones found in DBpedia

dataset, and each one contains structured information

extracted from Wikipedia. The DBpedia Lookup Service

can be used to look up DBpedia URIs by related key-

words.

In the educational context, tools are used to extract and

enrich entities found in educational forums [74]. Mudu et

al. [10] present a software architecture for linking the

SCORM standard to the LOD cloud. Zenaminer [10] al-

lows users to enrich the SCORM resources with com-

ments. Unstructured comments are automatically anno-

tated with DBpedia Spotlight linking them to the LOD

cloud through DBpedia. Chae et al. [75] present a system

that can help users to provide various types of multime-

dia data as linked data and considers RelFinder [5] as a

tool for exploration of the semantic relationships between

the multimedia resources.

Consumption may also happen directly inside applica-

tions, using APIs or SPARQL endpoints, for example, the

SPARQL endpoint for DBpedia

, which allows queries to

be performed against the data, taking advantage of con-

nections between different datasets.

Although the enrichment has been seen as an alterna-

tive in the educational domain, their costs are relatively

high. Enrichment typically requires finding suitable

Linked Data entities [76], but fully automatic annotations

are still challenging [33][35].

C. Semantic Web Browsers

Consumption of LD is discussed by Dadzie and Rowe

[77], the authors address the approaches and challenges

involved in data visualization. In this scenario, semantic

web browsers play an important role in supporting access

to data. The authors divide semantic browsers into two

types: text-based browsers and browsers offering data

visualization options.

Text-based browsers display the data using text struc-

tures such as tables and lists and may have more ad-

vanced features, allowing navigation through the data.

For example, semantic web browsers of this type include

Marbles

and URIburner

Browsers offering data visualization options use struc-

tures such as images, charts, graphs and timelines to pre-

sent the data. Examples include DBpedia Mobile

, IsaV-

, OpenLink Data Explorer (ODE)

and Tabulator

http://www.dbpedia-spotlight.org/

http://dbpedia.org/sparql

https://sourceforge.net/projects/marbles/files/marbles/Marbles%2

0Linked%20Data%20Engine%201.0/

http://linkeddata.uriburner.com/fct/

http://wiki.dbpedia.org/projects/dbpedia-mobile

https://www.w3.org/2001/11/IsaViz/ (latest version 2007)

http://ode.openlinksw.com/

https://www.w3.org/2005/ajar/tab.html

5. EDUCATIONAL DATASETS AND VOCABULARIES

5.1 Datasets

Datasets are particularly important for applications which

need to consume data and for publishers and data pro-

vider to connect your data with other possible datasets. In

the context of education, two types of datasets should be

considered [31]: (1) those directly related to educational

information, containing, for example, educational re-

sources, institutional data and educational indicators; and

(2) datasets from different domains (museums, agrono-

my, libraries, encyclopedias, etc.) which may be used in

educational settings.

The number of educational datasets is continuously

growing, making it difficult to provide an exhaustive

listing in the present study. In this article, datasets are

mentioned according to the higher number of citations in

the proposals analyzed and in the secondary studies re-

trieved by the search string. The list was also comple-

mented based on a search in the Open Data Inception

dataset repository.

The extraction of the main datasets used in the map-

ping studies provides support for answering the research

question (RQ02): “What are the Linked Data datasets used in

the context of Education?”

Based on analysis of the datasets, it was found that

there was significant use of datasets where the data is not

from a specific domain.

Some of the most cited datasets are DBpedia

, Free-

base

and GeoNames

. The DBpedia project [78] pro-

vides data automatically extracted from Wikipedia, using

structured information such as infobox tables, categoriza-

tion information, geo-coordinates and information from

external links. The DBpedia contains many links to other

dataset such as Wikidata, OpenCyc

, UMBEL

GeoNames, MusicBrainz

, UniProt

and Bio2RDF

Freebase is a Google Inc’s collaborative knowledge bases.

It was discontinued in 2010.However, the data was trans-

ferred to the collaborative knowledge base Wikimedia

Foundation's project called Wikidata

. GeoNames is a

geographical database containing over 8 million geo-

graphical names, including cities, towns, currencies,

mountains, etc. A comparison of some large datasets used

in LD can be found in [79].

The data in these datasets is used for: enriching educa-

tional content [10][12][80][81][82][83]; discovering new

information which can help educational practices

[22][84][26][24][85] and connecting local datasets to the

LOD cloud [86][87][88][89][90]. DBpedia, for instance, is

used in [73] to analyze universities ranking based on their

structured information. Robinson et al. [91] make use of

the categories provided by DBpedia to select the most

http://opendatainception.io/

http://wiki.dbpedia.org/

https://developers.google.com/freebase/

http://www.geonames.org/

http://sw.opencyc.org/

http://umbel.org/

https://musicbrainz.org/

http://www.uniprot.org/

http://bio2rdf.org/

https://www.wikidata.org/

information.

IEEE Transactions on Learning Technologies

8 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

suitable categories for describing learning objects.

Note that the most cited datasets are also known as

hubs in the LOD cloud, i.e., datasets that are interlinked

with others. Interlink datasets with DBpedia and other

hubs are essential to make a dataset findable to others.

Datasets from different domains which may be used in

educational settings include the dataset of TEDTalks

[92], which holds various conferences on a wide range of

topics, and data from museum collections (British Muse-

), cultural data (Europeana

), and libraries (Biblio-

thèque Nationale de France

Among the studies analyzed, some addressed the con-

struction and operation of datasets in the fields of medi-

cine, agriculture [93][94] and tourism [88]. Studies around

agriculture cited organic.edunet

, the NCBI Gene Expres-

sion Omnibus dataset

, Agris

, AGROVOC

, ASFA

and JITA

Datasets cited in the medical field included PubMed

and mEducator [62]. PubMed is a service of the US Na-

tional Library of Medicine which includes citations from

MEDLINE

and other scientific journals in life sciences.

The mEducator – Linked Educational Resources dataset is

intended to provide educational resources in a LD format.

Resources for mEducator are focused on the medical

field, covering content ranging from traditional teaching

to open learning, as well as experimental studies. (Answer

R4C2)

Several datasets used in Education were also frequent-

ly cited in the studies surveyed, such as Open University

[95]. Open University publishes information about the

courses it offers through its website and makes them

available in RDF format.

OCW Consortium is one of the initiatives in the global

promotion of free and unrestricted access to academic

knowledge. To solve the problem of standardization and

interoperability of the linked datasets made available,

some studies [24][96] using LD are underway to create an

infrastructure to integrate and access content across dif-

ferent repositories of universities associated with OCW.

Another project cited by several studies [56][97]

[98][99][100][101] is LinkedUp

, which maintains a cata-

log and a data repository, both of which are relevant and

useful in the educational setting. The objectives of the

LinkedUp Dataset Catalog (or Linked Education Cloud)

are to collect and make available all types of data sources

relevant for Education, to provide a shared resource, and

to develop the community interested in the Web of Data

for Education [40]. The datasets are obtained from various

sources, usually originating from a university which pub-

http://data.linkededucation.org/request/ted/sparql

http://collection.britishmuseum.org/

http://labs.europeana.eu/api/linked-open-data-SPARQL-endpoint

http://data.bnf.fr/

http://data.organic-edunet.eu/sparql

https://www.ncbi.nlm.nih.gov/gds

http://202.45.139.84:10035/catalogs/fao/repositories/agris

http://202.45.139.84:10035/catalogs/fao/repositories/agrovoc

http://www4.fao.org/asfa/asfa.htm

https://datahub.io/dataset/jita

http://pubmed.bio2rdf.org/sparql

https://www.nlm.nih.gov/bsd/pmresources.html

http://data.linkededucation.org/linkedup/catalog/

lishes its data, resource repositories or research. There are

also government datasets which contribute significantly

to linked data sets related to Education, for example,

generating statistics about the records and results of edu-

cational institutions.

Universities are also evolving considering the publica-

tion of their data, including administrative, academic,

scientific, and educational information, as well as other

data. These initiatives generate linked datasets containing

data which can be reused. These linked datasets include

the University of Southampton

, Greek University Open

Data

and the Linking Italian University Statistics Pro-

ject

The LAK dataset

provides a collection of data ex-

tracted from publications in the field of learning analytics

and educational data mining. It is regularly updated with

data from the LAK and EDM conferences.

Although there are a wide variety and number of

linked datasets available, they significantly vary regard-

ing correctness, accessibility, and quality [102], posing

some challenges on the reuse, retrieval, and linkage of

existing linked datasets. According to Auer et al. [103],

approx. 70% of linked datasets listed in DataHub present

issues, such as unresponsive SPARQL endpoints, lack of

descriptive metadata and outdated linked datasets, that

makes linked datasets inaccessible or unreliable.

Due to these problems, there is a demand for (automat-

ic) approaches that constantly check the availability, qual-

ity, and reliability of linked datasets. Moreover, addition-

al information about the status of the dataset, topics cov-

ered and the ephemerality of the information may be

considered to help data consumers and maintainers.

5.2 Vocabularies

Ontologies have become popular in the educational

field primarily because of what they promise: a shared

and common understanding between people and applica-

tion systems, reusable domain knowledge, enable explicit

domain assumptions, separate the domain knowledge

from the operational knowledge, and a comprehensive

analysis of domain knowledge [104].

Tiropanis et al. [105] state that the Linked Data move-

ment is related to an alternative approach to the develop-

ing of semantic technologies. Briefly, LD gives priority to

exposing data sources using lightweight vocabularies

first, to enable data interoperability and integration before

considering more complex ontologies in the context of

specific applications.

Reusing a vocabulary or ontology is equivalent to

adopting a standard, i.e., a common language which peo-

ple and machines can understand. This kind of standardi-

zation is crucial to improving discovery and reuse of

educational data and resources, as well as allowing in-

teroperability between repositories [4]. The use of com-

mon vocabularies to describe the subjects being dealt

with allows the contents of these repositories to be global-

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

http://www.auth.gr/sparql

http://sw.unime.it:8890/sparql

http://lak.linkededucation.org/

information.

IEEE Transactions on Learning Technologies

ly addressable, and to retrieve resources based primarily

on their relevance, without needing to consider their

origin [4].

The extraction of the main ontologies/vocabularies

used in the mapping studies provide support for answer-

ing the secondary question: “What ontologies and vocabu-

laries are being used in the publication of Linked Data in

the context of Education?” (SQ01).

The Dublin Core

is the most used vocabulary in the

analyzed proposals to describe of educational resources.

Other vocabularies for this same purpose are: LRMI

LOCO

, ALOCoM Core Ontology

(Educational re-

sources fragmentation) and Learning Object Metadata

Ontology (a mapping of IEEE LOM - Learning Object

Metadata

elements to RDF based on the Linked Data

principles).

Some proposals need to describe educational resources

of a specific domain, such as medicine, agriculture, arts,

chemistry or music. For this reason, many vocabularies

arise to represent characteristics of a domain, for example,

AGROVOC

, Body System

, BRO – Biomedical Resource

Ontology

, ChEBI

, Galen

, MeSH

, Music Ontology

PubChem

The AIISO

vocabulary provides classes and properties

to describe the internal organizational structure of educa-

tional institutions. Like AIISO are the Bowlogna ontolo-

, the Italian University Project (SWIUP) Ontology

and IntelLEO Organization Ontology

. Other vocabular-

ies and Ontologies can be found at Linked Open Vocabu-

laries (LOV)

, BioPortal

, JoinUp

, schema.org

and

https://linkededucation.wordpress.com/data-

models/schemas/.

Some recommendations are provided on the use of vo-

cabularies [4][35]: reuse of existing vocabularies, instead

of creating new ones; select vocabularies based on their

popularity, simplicity, consistency, scope, dataset com-

patibility and availability of documentation; and, where

necessary, create vocabulary mappings to reconcile terms.

Note that even though several educational datasets are

published following the LD principles, there are still a

few challenges concerning the vocabularies that may

hinder the creation of links inter datasets. For instance,

http://dublincore.org/specifications/

http://lrmi.net

http://jelenajovanovic.net/LOCO-Analyst/loco.html

http://jelenajovanovic.net/ontologies/loco/alocom-

core/spec/#overview

http://kmr.nada.kth.se/static/ims/md-lomrdf.html

http://aims.fao.org/vest-registry/vocabularies/agrovoc-

multilingual-agricultural-thesaurus

https://bioportal.bioontology.org/ontologies/ICD11-BODYSYSTEM

https://bioportal.bioontology.org/ontologies/BRO

http://www.ebi.ac.uk/chebi/

https://bioportal.bioontology.org/ontologies/GALEN

http://bioportal.bioontology.org/ontologies/MESH

http://musicontology.com/

https://pubchem.ncbi.nlm.nih.gov/rdf/rdf1.6.html

http://vocab.org/aiiso/schema

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

hema.rdf

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

http://bioportal.bioontology.org/

https://joinup.ec.europa.eu/catalogue/repository

http://schema.org/docs/schemas.html

D’Aquin et al [106] show that many linked data sets use

specific ad-hoc vocabularies that are hardly adopted by

others. Moreover, general use vocabularies, such as

FOAF, Dublin Core and SKOS, may not provide descrip-

tive information about the nature of the dataset. Thus, the

choice of vocabularies to be used to represent the dataset

must be carefully taken, otherwise it may make the reuse,

enrichment, and linkage of educational resources more

difficult. Although many efforts [104][107] [108][109] have

been devoted to the development and use of ontologies in

the educational field, there is no ontology that fully de-

scribes the different domains and pedagogical issues

related to teaching and learning process.

6. CHALLENGES

Dietze et al. [33] performed a study on the challenges and

approaches involved in connecting educational resources,

investigating the wealth of LD which already exists on

the Web. It deals with approaches for publishing and

connecting educational resources and data under the LD

principles, and the possibility of enriching their content

via connections between linked data sets.

The authors discuss four main challenges to achieving

interoperability of educational resources on the Web: (1)

integrating data distributed among heterogeneous educa-

tional repositories; (2) dealing with continuous changes

existing in the repositories; (3) mediating and transform-

ing metadata which describes resources; and (4) enriching

and interlinking unstructured metadata.

Regarding the challenges mentioned above, four ap-

proaches are also presented for building what the authors

call Linked Education

. These are: (P1) Principle of

Linked Data: applied to modeling and displaying

metadata of educational resources and educational ser-

vices and APIs; (P2) Services integration: heterogeneous

and distributed learning repositories i.e. their Web inter-

faces (services); (P3) Schema matching: the task of auto-

matically finding correspondences between elements or

concepts between two or more data models to allow the

interoperability between the wide variety of learning

repositories and property educational schemas available

on the Web. (P4) Data Interlinking, Clustering and En-

richment: automated enrichment and grouping mecha-

nisms are investigated to interlink the data produced by

(P3) with existing linked data sets as part of the LD cloud.

Many of the challenges mentioned in the studies analyzed

here are in-line with those presented by Dietze et al.[33].

These are common challenges in the use of Linked Data in

various areas. In this paper we expand the discussion of

these challenges by focusing on the educational context.

Some other challenges reported by the proposals ana-

lyzed were also mapped.

Even with the significant increase in educational re-

sources available on the Web, reuse is still a challenge.

According to Dietze et al. [110] some of reasons include:

the lack of quality of resources as well as their annota-

tions, diversity of metadata standards and vocabularies,

http://linkededucation.org

information.

IEEE Transactions on Learning Technologies

10 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

accessibility of data and the often poorly maintained

metadata descriptions, and raising concerns with respect

to trust and reliability when attempting to reuse third-

party data and resources.

A common problem in various areas is the difficulty in

defining what can / should be made available. There are

privacy and ownership issues regarding educational data,

with considerable concern on exposure of data, because

data linked with other data might reveal sensitive and

confidential information. As addressed in

[111][32][76][51], this question is still a challenge requir-

ing further discussion and a major commitment by vari-

ous stakeholders and sectors.

Although the privacy of educational data, e. g. person-

al data, academic performance, grades, etc. is a concern,

the availability of these data is extremely important to

understand the educational context, examine successful

and unsuccessful strategies and promote an innovative

environment in educational area. For this reason, it is

essential to design solutions that allow to publish educa-

tional data, to maintain the anonymity of individuals and

/ or to publish data from multiple users, making it more

difficult to infer individual characteristics.

The effort required to publish data as LD, as well as

the computational costs involved in its consumption, is a

problem discussed by Vega-Gorgojo et al. [35]. Depend-

ing on the volume of the knowledge base, the potential

for interactivity in LD may require a high computational

cost for data processing and availability. The authors

argue that although the impact of this problem is still

considerable in the education sector, it is not limited to

this context. Therefore, improvement can be expected

because of advances in LD discussions including, for

example, advances in developing practical methods,

standards, and tools to assist the publishing and interlink-

ing processes. Low quality in publication and interlinking

of educational data and resources is still an open issue,

discussed by several authors [18][111][76][112][113][51].

Dealing with inaccurate, incomplete and unexpected

information from the Web of Data becomes a problem

when there is no control in the publishing process and the

nature of the data is very diverse.

In the educational context, continuous changes are oc-

curring in repositories. As the context is constantly chang-

ing throughout the learning activities, Navarrete and

Lujan-Mora [34] explain that adjustments are needed

continuously to correctly represent the state of how peo-

ple, groups and activities are organized. Controlling such

changes and keeping them updated in the linked data sets

is a major challenge in using Linked Data in Education.

Considering the continuous changes occurring in re-

positories, Vasconcellos et al. [114] discussed automatic

ontology refinement as one possible way to allow updates

to ontologies so that they can reflect updates to the do-

main. Ontology refinement is the process which aims to

keep ontology coherent with the domain that it repre-

sents.

The problem of dataset unavailability and several oth-

er challenging issues are discussed in [115][116]. The

unavailability of linked datasets was a problem noticed

while performing the mapping in the present study, since

many linked datasets mentioned in the studies being

analyzed are no longer available. The most common sta-

tus codes include failures in the server to provide a valid

response such as ‘500 Internal Server Error’, ‘502 Bad

Gateway’, or ‘503 Service Unavailable’ and ‘404 Not

Found’.

Low quality of data results in additional complexity to

retrieving data even when it is openly available using

standards facilitating its discovery

[18][111][76][89][51][88].

Both the low quality of the data and the unavailability

of some linked datasets can have important impacts in the

educational context. The Learning Analytics area requires

on reliable and available information to analyze educa-

tional performance, generate indicators and make deci-

sions. Unavailability of a dataset or lack of update may,

for example, generate false analyzes.

The use of common vocabularies and completeness in

publication would be a way to alleviate the problems of

information retrieval, standardization in publishing and

interoperability between different educational reposito-

ries. However, the diversity of data in the context of edu-

cation often means that the vocabularies which are

known and available are not sufficient to ensure quality

in publication while also respecting the specificity of var-

ious subdomains in Education. Therefore, there is a need

for understanding the vocabularies available for Educa-

tion.

Another challenge in the use of vocabularies in the

context of Education [76] involves the need to build do-

main ontologies for different target domains. In addition,

it is important to make these vocabularies available in

service annotation knowledge bases, significantly im-

proving vocabulary reuse. Mapping between narrower

vocabularies and broader, more widely used, vocabular-

ies is also discussed as an alternative to address the spe-

cifics of educational data.

The use of metadata standards in educational resource

repositories could be a facilitator in providing educational

resources as Linked Data and ensuring interoperability of

repositories. However, certain challenges remain, as dis-

cussed in [117][70][89]. While the main repositories have

metadata standards (Learning Object Metadata, Dublin

Core etc.), other repositories describe their resources us-

ing their methods, such as XML-based schemas and het-

erogeneous taxonomies. Consequently, the transfor-

mation of schema and data to RDF is a continuing prob-

lem. Thus, the translation of existing metadata standards

to RDF vocabularies is one of the main challenges.

The whole argument regarding new knowledge dis-

covery, interoperability between repositories and enrich-

ment of educational resources is supported by the possi-

bility of linking data from different linked data sets. The

data interlinking step also presents various challenges

which are related, for example, to identifying correspond-

ing instances, discovering semantic relations between

data, scalability, and need for developments in tools and

practices. These are challenges that still need to be ad-

dressed in the LD area. The evolution of research will

information.

IEEE Transactions on Learning Technologies

undoubtedly have an impact in the context of Education.

A far-ranging discussion regarding the challenges in-

volved in data interlinking can be found in [67][65]. As-

pects which are more specific to the context of Education

are addressed in [89][52]. The authors explain that the

existing metadata for educational resources is usually

provided based on informal and unstructured data, and

the use of controlled vocabularies is limited and frag-

mented. To enable machine processing and interoperabil-

ity, educational metadata needs to be enriched, trans-

forming it into structured and formal descriptions, and

linking it to LD vocabularies and linked datasets widely

established on the Web.

7. FINAL CONSIDERATIONS

The LD area is constantly evolving and many challenges

continue to motivate the development of research in this

field. The proposals analyzed in this work showed that

there is an effort by different institutions to apply the LD

principles and technologies to solve known problems

around Educational Technologies, such as interoperabil-

ity of educational data and resources, access to data for

various analyses, enrichment of material, and recommen-

dation and personalization content.

The study conducted in the present work has aimed to

provide a mapping of the Linked Data research scenario

within the context of Education. For this purpose, this

literature review has attempted to identify the proposals,

the objectives for using LD, the tools, linked datasets and

vocabularies used on LD in Education. The primary goal

was to provide an overview of how this topic is being

investigated, while also guiding future decisions on pro-

jects seeking to adopt LD in Education. Analyzing a larg-

er volume of papers (114 out of 1,384 pre-selected papers)

also made it possible to obtain evidence regarding how

the results of the recent researches in Semantic Technolo-

gies in education, more specifically the movement of

linked data, are being applied in practice. Thus, it is pos-

sible to understand which approaches, technologies,

tools, ontologies and linked datasets are being adopted

and what challenges are encountered by the projects.

In addition, the present study has identified challenges

currently faced in each stage of the process, while also

providing an overview of the current landscape of the

area, which can serve to guide future work and suggest

possible avenues for further research.

ACKNOWLEDGMENT

This work was partially supported by FAPERJ (through

grant E-26-102.256/2013 - BBP/Bursary Associa: Explor-

ing a Semantic and Social Learning-Teaching Environ-

ment), CNPQ (project: 312039/2015-8 – DT/Bursary Inte-

grating Pedagogical Practices and Methods and Tools of

Educational Data Analysis) and the National Institute of

Science and Technology (INCT) on Web Science.

REFERENCES

[1] C. Bizer, T. Heath, and T. Berners-lee, “Linked Data - The

Story So Far,” Int. J. Semant. Web Inf. Syst., vol. 5, no. 3, pp.

1–22, 2009.

[2] T. Berners-Lee, “Linked Data,”

http://www.w3.org/DesignIssues/LinkedData.html, 2006.

[Online]. Available:

http://www.w3.org/DesignIssues/LinkedData.html.

[3] O. Lassila and Ralph R. Swick, “Resource Description

Framework \uppercase{(RDF)} Model and Syntax

Specification,” no. February, 1999.

[4] M. D’Aquin, “Linked Data for Open and Distance

Learning,” 2012.

[5] F. Zablith, M. Fernandez, and M. Rowe, “Production and

consumption of university Linked Data,” Interact. Learn.

Environ., vol. 23, no. 1, pp. 55–78, 2015.

[6] S. Kagemann, “MOOCLink : Building and Utilizing Linked

Data from Massive Open Online Courses,” pp. 373–380,

2015.

[7] G. Pirrotta, “Linking Italian university statistics,” Proc. 6th

Int. Conf. Semant. Syst. - I-SEMANTICS ’10, p. 1, 2010.

[8] G. G. Juanes, A. R. Barrios, J. L. R. García, L. G. Medina, R.

D. Adán, and P. G. Yanes, “Linked data strategy to achieve

interoperability in higher education,” WEBIST 2014 - Proc.

10th Int. Conf. Web Inf. Syst. Technol., vol. 2, pp. 57–64, 2014.

[9] F. Zablith, “Interconnecting and Enriching Higher

Education Programs Using Linked Data,” Proc. 24th Int.

Conf. World Wide Web, pp. 711–716, 2015.

[10] E. Mudu, L. Schiatti, G. Rizzo, and A. Servetti, “Zenaminer:

Driving the SCORM standard towards the web of data,”

CEUR Workshop Proc., vol. 717, pp. 1–15, 2011.

[11] N. Piedra and J. Chicaiza, “Using linked open data to

improve the search of open educational resources for

engineering students,” pp. 6–8, 2013.

[12] H. Q. Yu, C. Pedrinaci, S. Dietze, and J. Domingue, “Using

linked data to annotate and search educational video

resources for supporting distance learning,” IEEE Trans.

Learn. Technol., vol. 5, no. 2, pp. 130–142, 2012.

[13] G. Gutiérrez-carreón, T. Daradoumis, and J. Jorba,

“Integrating Learning Services in the Cloud : An Approach

that Benefits Both Systems and Learning,” Educ. Technol.

Soc., vol. 18, pp. 145–157, 2015.

[14] Adolfo Ruiz-Calleja, G. Vega-Gorgojo, J. I. Asensio-Pérez,

M. L. Bote-Lorenzo, E. Gómez-Sánchez, and C. Alario-

Hoyos, “A Linked Data approach for the discovery of

educational ICT tools in the Web of Data,” Comput. Educ.,

vol. 59, no. 3, pp. 952–962, 2012.

[15] E. Kaldoudi, N. Dovrolis, D. Giordano, S. Dietze, and M.

Keynes, “Educational Resources as Social Objects in

Semantic Social Networks,” 2011.

[16] C. Kiourt, A. Koutsoudis, F. Arnaoutoglou, G. Petsa, S.

Markantonatou, and G. Pavlidis, “A dynamic web-based

3D virtual museum framework based on open data,” 2015

Digit. Herit., vol. 2, pp. 647–650, 2015.

[17] M. Fernandez, M. D’Aquin, and E. Motta, “Linking data

across universities: An integrated video lectures dataset,”

Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif.

Intell. Lect. Notes Bioinformatics), vol. 7032 LNCS, no. PART

2, pp. 49–64, 2011.

[18] E. Otero-García, J. C. Vidal, M. Lama, A. Bugarín, and J.

Domenech, “A context-based algorithm for annotating

educational content with Linked Data,” in 1st Workshop on

Mining the Future Internet, 2010, vol. 685.

[19] G. Á. Rey, I. Celino, P. Alexopoulos, D. Damljanovic, M.

Damova, N. Li, and V. Devedzic, “Semi-automatic

generation of quizzes and learning artifacts from linked

data,” CEUR Workshop Proc., vol. 840, pp. 6–8, 2012.

[20] A. Carbonaro, “Interlinking e-learning resources and the

web of data for improving student experience,” J. E-

Learning Knowl. Soc., vol. 8, no. 2, pp. 33–44, 2012.

[21] J. Y. Wu and C. Le Wu, “The Study of User Model of

Personalized Recommendation System Based on Linked

information.

IEEE Transactions on Learning Technologies

12 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

Course Data,” Appl. Mech. Mater., vol. 519–520, pp. 1609–

1612, Feb. 2014.

[22] J. Chicaiza, N. Piedra, J. Lopez, and E. T. Caro, “Promotion

of self-learning by means of Open Educational Resources

and semantic technologies,” 2015 Int. Conf. Inf. Technol.

Based High. Educ. Training, ITHET 2015, pp. 1–6, 2015.

[23] S. Softic, L. De Vocht, B. Taraghi, M. Ebner, E. Mannens,

and R. V De Walle, “Leveraging Learning Analytics in a

Personal Learning Environment using Linked Data,” Bull.

IEEE Tech. Comm. Learn. Technol., vol. 16, no. 4, pp. 10–13,

2014.

[24] N. Piedra, J. López, C. Jorge, and E. Tovar, “Seeking Open

Educational Resources to Compose Massive Open Online

Courses in Engineering Education An Approach Based on

Linked Open Data,” J. Univers. Comput. Sci., vol. 21, no. 5,

pp. 679–711, 2015.

[25] M. Zotou, A. Papantoniou, K. Kremer, V. Peristeras, and E.

Tambouris, “Implementing ‘ Rethinking Education ’:

Matching Skills Profiles with Open Courses through Linked

Open Data technologies,” vol. 16, no. 4, 2014.

[26] J. Lopez-Vargas, N. Piedra, J. Chicaiza, and E. Tovar, “OER

Recommendation for Entrepreneurship Using a Framework

Based on Social Network Analysis,” Rev. Iberoam. Tecnol. del

Aprendiz., vol. 10, no. 4, pp. 262–268, 2015.

[27] N. Piedra, J. Chicaiza, J. López, and E. Tovar Caro,

“Supporting openness of MOOCs contents through of an

OER and OCW framework based on Linked Data

technologies,” in IEEE Global Engineering Education

Conference, EDUCON, 2014, pp. 1112–1117.

[28] D. Taibi and S. Dietze, “Fostering analytics on learning

analytics research: The LAK dataset,” CEUR Workshop Proc.,

vol. 974, pp. 5–7, 2013.

[29] R. A. Maturana, M. E. Alvarado, S. Lopez-Sola, M. J. Ibanez,

and L. R. Elosegui, “Linked Data based applications for

Learning Analytics Research: Faceted searches, enriched

contexts, graph browsing and dynamic graphic

visualisation of data,” CEUR Workshop Proc., vol. 974, 2013.

[30] B. E. Penteado, “Correlational Analysis Between School

Performance and Municipal Indicators in Brazil Supported

by Linked Open data,” pp. 507–512, 2016.

[31] C. Keßler, M. D’Aquin, and S. Dietze, “Linked Data for

Science and Education,” Semant. Web, vol. 1, pp. 1–5, 2009.

[32] K. Krieger and D. Rösner, “Linked Data in E-Learning: A

Survey,” Semant. Web 0, vol. 0, pp. 1–9, 2011.

[33] S. Dietze, S. Sanchez-Alonso, H. Ebner, H. Q. Yu, D.

Giordano, I. Marenzi, and B. P. Nunes, “Interlinking

educational resources and the web of data: A survey of

challenges and approaches,” Program, vol. 47, no. 1, pp. 60–

91, 2013.

[34] R. Navarrete and S. Lujan-Mora, “Use of Linked Data to

enhance Open Educational Resources,” Inf. Technol. Based

High. Educ. Train. (ITHET), 2015 Int. Conf. on. IEEE, pp. 1–6,

2015.

[35] G. Vega-Gorgojo, J. I. Asensio-Pérez, E. Gómez-Sánchez, M.

L. Bote-Lorenz, J. A. Muñoz-Cristóbal, and A. Ruiz-Calleja,

“A review of linked data proposals in the learning

domain,” J. Univers. Comput. Sci., vol. 21, no. 2, pp. 326–364,

2015.

[36] B. Kitchenham and S. Charters, “Guidelines for performing

Systematic Literature Reviews in Software Engineering,”

Engineering, vol. 2, p. 1051, 2007.

[37] J. Biolchini, P. G. Mian, A. Candida, and C. Natali,

“Systematic Review in Software Engineering,” Engineering,

vol. 679, no. May, pp. 165–176, 2005.

[38] H. Qing, S. Dietze, D. Giordano, D. Taibi, E. Kaldoudi, and

N. Dovrolis, “Linked education : interlinking educational

resources and the web of data,” 27th Annu. ACM Symp.

Appl. Comput., pp. 366–371, 2012.

[39] D. Mouromtsev and M. D’Aquin, Eds., Open Data for

Education: Linked, Shared, and Reusable Data for Teaching and

Learning, vol. 9500. Cham: Springer International

Publishing, 2016.

[40] M. D’Aquin, “On the Use of Linked Open Data in

Education: Current and Future Practices,” Lect. Notes

Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect.

Notes Bioinformatics), vol. 9500, pp. 153–165, 2016.

[41] N. Piedra, J. Chicaiza, J. López, and E. Tovar, “Using linked

open data to improve the search of open educational

resources for engineering students,” Front. Educ. Conf., pp.

558–560, 2013.

[42] F. Zablith, M. D’Aquin, S. Brown, and L. Green-Hughes,

“The Open University ’ s repository of research

publications Consuming Linked Data within a Large

Educational Organization Conference Item,” in Second

International Workshop on Consuming Linked Data (COLD) at

10th International SemanticWeb Conference, 2011, pp. 23–27.

[43] C. Bizer and R. Cyganiak, “D2RQ — Lessons Learned,” in

W3C Workshop on RDF Access to Relational Databases, 2007,

pp. 1–10.

[44] E. Rajabi, M.-A. Sicilia, and S. Sanchez-Alonso,

“Discovering duplicate and related resources using an

interlinking approach: The case of educational datasets,” J.

Intell. Mater. Syst. Struct., vol. 26, no. 5, pp. 599–613, 2015.

[45] N. Dovrolis, T. Stefanut, S. Dietze, H. Q. Yu, C. Valentine,

and E. Kaldoudi, “Semantic annotation and linking of

medical educational resources,” IFMBE Proc., vol. 37, pp.

1400–1403, 2011.

[46] S. S. Sahoo, W. Halb, S. Hellmann, K. Idehen, T. Thibodeau

Jr, S. Auer, J. Sequeda, and A. Ezzat, “A survey of Current

approaches for mapping of relational databases to RDF,”

Web Semant. Sci. Serv. Agents World Wide Web, vol. 6, no. 1,

p. 15, 2009.

[47] S. Das, S. Sundara, and R. Cyganiak, “R2RML: RDB to RDF

Mapping Language,” W3C, 2012. [Online]. Available:

https://www.w3.org/TR/2012/PR-r2rml-20120814/.

[48] N. Thangsupachai, S. Niwattanakul, and N. Chamnongsri,

“Learning Object Metadata Mapping for Linked Open

Data,” Emerg. Digit. Libr. - Res. Pract., vol. 8839, no. 2, pp.

122–129, 2014.

[49] K. Rohloff, M. Dean, I. Emmons, D. Ryder, and J. Sumner,

“An Evaluation of Triple-Store Technologies for Large Data

Stores,” Move to Meaningful Internet Syst. 2007 OTM 2007

Work., vol. 4806, pp. 1105–1114, 2007.

[50] W3C, “LargeTripleStores,” 2016. [Online]. Available:

https://www.w3.org/wiki/LargeTripleStores. [Accessed:

28-Nov-2016].

[51] W. Alcantara, J. Bandeira, A. Barbosa, and A. Lima,

“Desafios no uso de Dados Abertos Conectados na

Educação Brasileira,” An. do 4o DesafIE - Work. Desafios da

Comput. Apl. à Educ., 2015.

[52] S. Dietze, H. E. Salvador Sanchez-Alonso, H. Q. Yu, D.

Giordano, I. Marenzi, and B. P. Nunes, “Interlinking

educational resources and the web of data: A survey of

challenges and approaches,” Emerald insight, vol. 29, no. 5,

pp. 494–519, 2009.

[53] A. Zouaq, J. Jovanovic, S. Joksimovic, and D. Gasevic,

“Linked Data for Learning Analytics: Potentials and

Challenges,” Handb. Learn. Anal., pp. 347–355, 2017.

[54] D. Taibi, B. Fetahu, and S. Dietze, “Towards integration of

web data into a coherent educational data graph,” Proc.

22nd Int. Conf. World Wide Web companion, pp. 419–424, 2013.

[55] H. Q. Yua, S. Dietzeb, D. Taibic, D. Giordanod, E.

Kaldoudie, and J. Domingue, “A Linked Dataset of Medical

Educational Resources,” Br. J. Educ. Technol., vol. 46, no. 5,

pp. 1123–1129, 2015.

[56] H. Hajri, Y. Bourda, and F. Popineau, “Querying

Repositories of OER Descriptions: The Challenge of

Educational Metadata Schemas Diversity,” Des. Teach.

Learn. a Networked World, vol. 9307, pp. 582–586, 2015.

[57] X. Ma, “Knowledge point-based approach to interlink open

information.

IEEE Transactions on Learning Technologies

education resources,” Lect. Notes Comput. Sci. (including

Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol.

7882 LNCS, pp. 717–721, 2013.

[58] E. Rajabi, M.-Á. Sicilia, and S. Sánchez-Alonso,

“Interlinking Educational Data : An Experiment with

GLOBE Resources,” TEEM ’13 Proc. First Int. Conf. Technol.

Ecosyst. Enhancing Multicult., pp. 339–348, 2013.

[59] M. Hausenblas, W. Halb, and Y. Raimond, “Scripting user

contributed interlinking,” CEUR Workshop Proc., vol. 368,

2008.

[60] F. Scharffe, Y. Liu, and C. Zhou, “Rdf-ai: an architecture for

rdf datasets matching, fusion and interlink,” Proc. IJCAI

2009 Work. Identity, Ref. …, 2009.

[61] J. Lehmann, S. Auer, S. Capadisli, K. Janowicz, C. Bizer, T.

Heath, A. Hogan, and T. Berners-Lee, “LDOW2017: 10th

Workshop on Linked Data on the Web,” Proc. 26th Int. Conf.

World Wide Web Companion, pp. 1679–1680, 2017.

[62] S. T. Konstantinidis, L. Ioannidis, D. Spachos, C. Bratsas,

and P. D. Bamidis, “MEducator 3.0: Combining semantic

and social web approaches in sharing and retrieving

medical education resources,” Proc. - 2012 7th Int. Work.

Semant. Soc. Media Adapt. Pers. SMAP 2012, pp. 42–47, Dec.

2012.

[63] L. A. P. P. Leme, G. R. Lopes, B. P. Nunes, M. A. Casanova,

and S. Dietze, “Identifying candidate datasets for data

interlinking,” Int. Conf. Web Eng., vol. 7977 LNCS, pp. 354–

366, 2013.

[64] G. R. Lopes, L. Andr, G. Rabello Lopes, L. Paes Leme, B.

Pereira Nunes, M. Casanova, and S. Dietze, “Two

Approaches to the Dataset Interlinking Recommendation

Problem,” Web Inf. Syst. Eng. – WISE 2014 SE - 25, vol. 8786,

pp. 324–339, 2014.

[65] S. Wölger, C. Hofer, K. Siorpaes, S. Thaler, E. Simperl, and

T. Bürger, “Interlinking data-approaches and tools,”

Innsbruck - Austria, 2011.

[66] E. Rajabi, M.-A. Sicilia, and S. Sanchez-Alonso, “An

empirical study on the evaluation of interlinking tools on

the Web of Data,” J. Inf. Sci., vol. 40, no. 5, pp. 637–648, Oct.

2014.

[67] K. Nguyen, R. Ichise, and B. Le, “Interlinking Linked Data

Sources Using a Domain-Independent System,” in Joint

International Semantic Technology Conference, 2013, pp. 113–

128.

[68] L. von Ahn, “Games with a Purpose,” Computer (Long.

Beach. Calif)., vol. 39, no. 6, pp. 92–94, 2006.

[69] O. Hassanzadeh, R. Xin, R. J. Miller, A. Kementsietsidis, L.

Lim, and M. Wang, “Linkage Query Writer,” Proc. VLDB

Endow., vol. 2, no. 2, pp. 1590–1593, 2009.

[70] F. Zablith, M. D’Aquin, S. Brown, and L. Green-Hughes,

“Consuming Linked Data within a large educational

organization,” CEUR Workshop Proc., vol. 782, 2011.

[71] M. Foulonneau, “Generating educational assessment items

from linked open data: The case of DBpedia,” Lect. Notes

Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect.

Notes Bioinformatics), vol. 7117 LNCS, pp. 16–27, 2012.

[72] V. Uren, P. Cimiano, J. Iria, S. Handschuh, M. Vargas-vera,

E. Motta, and F. Ciravegna, “Semantic annotation for

knowledge management: Requirements and a survey of the

state of the art,” Web Semant. Sci. Serv. Agents World Wide

Web, vol. 4, no. 1, pp. 14–28, 2006.

[73] R. Meymandpour and J. G. Davis, “Ranking Universities

Using Linked Open Data,” J. Stud. Int. Educ., vol. 18, no. 2,

pp. 318–327, 2007.

[74] B. P. Nunes, R. Kawase, B. Fetahu, M. A. Casanova, and G.

H. B. de Campos, “Educational forums at a glance: Topic

extraction and selection,” Lect. Notes Comput. Sci. (including

Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol.

8787, pp. 351–364, 2014.

[75] J. Chae, Y. Cho, M. Lee, S. Lee, M. Choi, and S. Park,

“Design and implementation of a system for creating

multimedia linked data and its applications in education,”

Multimed. Tools Appl., 2015.

[76] Z. Jeremić, J. Jovanović, and D. Gašević, “Personal Learning

Environments on the Social Semantic Web,” Semant. web,

vol. 4, no. 1, pp. 23–51, 2013.

[77] A. S. Dadzie and M. Rowe, “Approaches to visualising

Linked Data: A survey,” Semant. Web, vol. 2, no. 2, pp. 89–

124, 2011.

[78] S. Auer, C. Bizer, G. Kobilarov, J. Lehmann, R. Cyganiak,

and Z. Ives, “DBpedia: A Nucleus for a Web of Open Data,”

Semant. Web, vol. 4825, pp. 722–735, 2007.

[79] M. Färber, B. Ell, C. Menne, and A. Rettinger, “A

Comparative Survey of DBpedia , Freebase ,” vol. 1, pp. 1–

5, 2015.

[80] V. Vasiliev, F. Kozlov, D. Mouromtsev, S. Stafeev, and O.

Parkhimovich, “ECOLE: An Ontology-Based Open Online

Course Platform,” in Lecture Notes in Computer Science

(including subseries Lecture Notes in Artificial Intelligence and

Lecture Notes in Bioinformatics), vol. 9500, 2016, pp. 41–66.

[81] T. Kelkar, A. Ray, and V. Choppella, “SangeetKosh: An

open web platform for music education,” Proc. - IEEE 15th

Int. Conf. Adv. Learn. Technol. Adv. Technol. Support. Open

Access to Form. Informal Learn. ICALT 2015, pp. 5–9, 2015.

[82] F. Moritz, M. Siebert, and C. Meinel, “Improving Search in

Tele-Lecturing: Using Folksonomies as Trigger to Query

Semantic Datasets to extract additional metadata,” Proc. Int.

Conf. Web Intell. Min. Semant. - WIMS ’11, 2011.

[83] J. Robinson, J. Stan, and M. Ribière, “Using linked data to

reduce learning latency for e-book readers,” Lect. Notes

Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect.

Notes Bioinformatics), vol. 7117 LNCS, no. January, pp. 28–

34, 2012.

[84] C. Daraio and A. Bonaccorsi, “Beyond university rankings?

Generating new indicators on universities by linking data

in open platforms,” J. Assoc. Inf. Sci. Technol., vol. 14, no. 4,

p. n/a-n/a, Mar. 2016.

[85] J. Y. Wu and C. L. Wu, “The study of user model of

personalized recommendation system based on linked

course data,” Appl. Mech. Mater., vol. 519–520, no.

Computer and Information Technology, p. 1607, 2014.

[86] G. Kobilarov, T. Scott, Y. Raimond, S. Oliver, C. Sizemore,

M. Smethurst, C. Bizer, and R. Lee, “Media meets semantic

web - How the bbc uses dbpedia and linked data to make

connections,” Lect. Notes Comput. Sci. (including Subser. Lect.

Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 5554 LNCS,

pp. 723–737, 2009.

[87] J. Chicaiza, N. Piedra, J. López-Vargas, and Tovar-

Edmundo, “Domain Categorization of Open Educational

Resources Based on Linked Data,” Knowl. Eng. Semant. Web,

pp. 15–28, 2014.

[88] A. M. Fermoso, M. Mateos, M. E. Beato, and R. Berjón,

“Open linked data and mobile devices as e-tourism tools. A

practical approach to collaborative e-learning,” Comput.

Human Behav., vol. 51, pp. 618–626, 2015.

[89] S. Dietze, E. Kaldoudi, N. Dovrolis, D. Giordano, C.

Spampinato, M. Hendrix, A. Protopsaltis, D. Taibi, and H.

Q. Yu, “Socio-semantic integration of educational resources

- The case of the mEducator project,” J. Univers. Comput.

Sci., vol. 19, no. 11, pp. 1543–1569, 2013.

[90] D. Hladky, “LOD Russia – Enabling Russian National

Knowledge with Scientific Open Data Research challenges,”

pp. 1–5, 2011.

[91] M. Lama, J. C. Vidal, E. Otero-Garc??a, A. Bugar??n, and S.

Barro, “Semantic linking of learning object repositories to

DBpedia,” Educ. Technol. Soc., vol. 15, no. 4, pp. 47–61, 2012.

[92] D. Taibi, S. Chawla, S. Dietze, I. Marenzi, and B. Fetahu,

“Exploring TED talks as linked data for education,” Br. J.

Educ. Technol., vol. 46, no. 5, pp. 1092–1096, 2015.

[93] E. Rajabi, S. Sanchez-Alonso, M. A. Sicilia, and N.

Manouselis, “A linked and open dataset from a network of

information.

IEEE Transactions on Learning Technologies

14 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

learning repositories on organic agriculture,” Br. J. Educ.

Technol., 2015.

[94] M.-A. Sicilia, H. Ebner, S. Sánchez-Alonso, A. Abián, and E.

García-Barriocanal, “Navigating learning resources through

linked data: a preliminary report on the re-design of

Organic . Edunet,” Proc. Linked Learn., vol. 2011, p. 1st, 2011.

[95] H. Qing, S. Dietze, D. Giordano, D. Taibi, E. Kaldoudi, and

N. Dovrolis, “The Open University ’ s repository of

research publications Linked education : interlinking

educational resources and the web of data,” in The 27th

ACM Symposium On Applied Computing (SAC-2012), Special

Track on Semantic Web and Applications, 2012.

[96] N. Piedra, E. Tovar, R. Colomo-Palacios, J. Lopez-Vargas,

and J. Alexandra Chicaiza, “Consuming and producing

linked open data: the case of OpenCourseWare,” Progr.

Electron. Libr. Inf. Syst., vol. 48, no. 1, pp. 16–40, 2014.

[97] D. Taibi, S. Dietze, B. Fetahu, and G. Fulantelli, “Exploring

type-specific topic profiles of datasets : a demo for

educational linked data,” 2011.

[98] M. Coccoli and I. Torre, “Interacting with annotated objects

in a semantic web of things application,” J. Vis. Lang.

Comput., vol. 25, no. 6, pp. 1012–1020, 2014.

[99] A. Mikroyannidis and J. Domingue, “Interactive learning

resources and linked data for online scientific

experimentation,” Proc. 22nd Int. Conf. World Wide Web

companion, pp. 431–434, 2013.

[100] S. Zerr, M. D’Aquin, I. Marenzi, D. Taibi, A. Adamou, and

S. Dietze, “Towards analytics and collaborative exploration

of social and linked media for technology-enchanced

learning scenarios,” CEUR Workshop Proc., vol. 1151, 2014.

[101] M. Coccoli and I. Torre, “Interaction with objects and

objects annotation in the semantic web of things,” Proc.

DMS 2014 - 20th Int. Conf. Distrib. Multimed. Syst., pp. 383–

390, 2014.

[102] C. Guéret, P. Groth, C. Stadler, and J. Lehmann, “Assessing

linked data mappings using network measures,” Simperl,

E., Cimiano, P., Polleres, A., Corcho, O., Presutti, V., vol. 7295

LNCS, pp. 87–102, 2012.

[103] S. Auer, I. Ermilov, J. Lehmann, and M. Martin,

“LODStats.” [Online]. Available: http://stats.lod2.eu/.

[Accessed: 26-Apr-2017].

[104] D. Kanellopoulos, “Ontology-based Learning Applications:

A Development Methodology.,” … Conf. Softw. …, no.

JANUARY, 2006.

[105] T. Tiropanis, H. C. Davis, S. A. Cerri, T. Tiropanis, H. C.

Davis, S. A. Cerri, and S. Technologies, “Semantic

Technologies and Learning,” pp. 3029–3032, 2012.

[106] M. d’Aquin, A. Adamou, and S. Dietze, “Assessing the

educational linked data landscape,” Proc. 5th Annu. ACM

Web Sci. Conf. - WebSci ’13, pp. 43–46, 2013.

[107] P. MULHOLLAND, Z. ZDRAHAL, J. DOMINGUE, M.

HATALA, and A. BERNARDI, “A methodological

approach to supporting organizational learning,” Int. J.

Hum. Comput. Stud., vol. 55, no. 3, pp. 337–367, 2001.

[108] R. Mizoguchi and G. H. Canada, “Using Ontological

Engineering to Overcome Common AI-ED Problems,” J.

Artif. Intell. Educ., vol. 11, pp. 107–121, 2000.

[109] V. Devedzi, “Education and the Semantic Web,” Int. J. Artif.

Intell. Educ., vol. 14, pp. 39–65, 2004.

[110] S. Dietze, D. Taibi, and P. Barker, “Analysing and

Improving Embedded Markup of Learning Resources on

the Web,” pp. 283–292, 2017.

[111] M. Svensson, A. Kurti, and M. Milrad, “Enhancing

emerging learning objects with contextual metadata using

the linked data approach,” 6th IEEE Int. Conf. Wireless, Mob.

Ubiquitous Technol. Educ., pp. 50–56, Apr. 2010.

[112] J. C. Vidal, M. Lama, E. Otero-García, and A. Bugarín,

“Graph-based semantic annotation for enriching

educational content with linked data,” Knowledge-Based

Syst., vol. 55, pp. 29–42, 2014.

[113] N. Bassiliades, “Collecting University Rankings for

Comparison Using Web Extraction and Entity Linking

Techniques,” Commun. Comput. Inf. Sci., vol. 469, pp. 23–46,

2014.

[114] S. Vasconcellos, K. Revoredo, and F. Bai, “How Can

Ontology Design Patterns Help Ontology Refinement ?,”

vol. 12, pp. 4–16, 2014.

[115] A. Hogan, P. Hitzler, and K. Janowicz, “Linked Dataset

Description Papers at the Semantic Web Journal : A Critical

Assessment,” vol. 7, pp. 1–3, 2016.

[116] F. Radulovic, N. Mihindukulasooriya, R. García-Castro, and

A. Gómez Pérez, “A comprehensive quality model for

Linked Data,” Undefined, vol. 1, no. 0, pp. 1–5, 2015.

[117] H. Q. Yu, S. Dietze, N. Li, C. Pedrinaci, D. Taibi, N.

Dovrolls, T. Stefanut, E. Kaldoudi, and J. Domingue, “The

Open University ’ s repository of research publications A

linked data-driven & service-oriented architecture for

sharing educational resources,” 1st Int. Work. eLearning

Approaches Linked Data Age (Linked Learn. 2011), 8th Ext.

Semant. Web Conf., 2011.

Crystiam Kelle Pereira e Silva

http://lattes.cnpq.br/1675106850769427

Crystiam is PhD Candidate at the

Graduate Program of Informatics at the

Federal University of the State of Rio

de Janeiro (UNIRIO). She also works as

system analyst at the Federal

University of Juiz de Fora, UFJF. Her

main interests lie in the Semantic Web

and Distance Learning fields.

Sean Wolfgand Matsui Siqueira

http://lattes.cnpq.br/2562652838103607

DSc in Computer Science, Sean is an

associate professor at the Federal

University of the State of Rio de

Janeiro (UNIRIO). He is editor-in-chief

of the RBIE: Brazilian Journal on

Computers in Education and co-editor

of the iSYS: Brazilian Journal of

Information Systems. He is the

founder and coordinator of the

Semantics and Learning research

group and is a member of the special committees on

Computers and Education (CEIE) and on Information

Systems (CESI), both from the Brazilian Computer Society

(SBC). His research interests are knowledge

representation, web science (including social and

semantic web) and advanced technologies for teaching

and learning.

Bernardo Pereira Nunes

http://lattes.cnpq.br/1728746187630338

DSc in Computer Science, Bernardo

is researcher and adjunct professor

at the Pontifical University Catholic

of Rio de Janeiro (PUC-Rio) and

collaborator at the Graduate

Program of Informatics at the

Federal University of the State of

Rio de Janeiro (UNIRIO). He also

information.

IEEE Transactions on Learning Technologies

works as senior systems analyst at the Center of Distance

Learning, PUC-Rio and is associate editor of the RBIE:

Brazilian Journal on Computers in Education. He has

experience in Computer Science, focusing on Semantic

Web, Web Science and Distance Learning.

Stefan Dietze

https://stefandietze.wordpress.com/

Research group leader at the L3S

Research Center of the Leibniz

University Hanover, Germany,

following previous positions at the Knowledge Media

Institute (KMI) of The Open University (UK) and the

Fraunhofer Institute for Software and Systems

Engineering (ISST, now part of Fraunhofer FOKUS).

Stefan’s interests are in the fields of web & data science,

artificial intelligence and knowledge engineering &

discovery, and in particular, the extraction and fusion of

knowledge from heterogeneous Web data in a variety of

application domain.

information.

IEEE Transactions on Learning Technologies

16 IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID

TRENDS AND APPLICATIONS OF BIG DATA IN EDUCATION

Article

Full-text available

Jun 2023

Big data technologies have facilitated innovation across various industries, including healthcare, technology, and education. The utilization of this innovation is now required within the education sector at all levels. It is a progressing framework named as "Education 4.0" to accommodate the numerous needs of this sector. Specifically, this article first describes the architecture of big data and its properties compatible to Education sector. The process of utilizing large educational data and data mining techniques in education are explained. To find solutions to diverse educational problems, an ever-increasing variety of mining tools have been assembled. This study also provides insights into these techniques and other data sources for gathering educational data. Several applications have been developed to promote the use of big data techniques in higher education and highlight its benefits for the sector, but there are still a number of issues that need to be considered. We will highlight the obstacles of using big data techniques in higher education.

Knowledge Graphs: Opportunities and Challenges

Preprint

Full-text available

Mar 2023

With the explosive growth of artificial intelligence (AI) and big data, it has become vitally important to organize and represent the enormous volume of knowledge appropriately. As graph data, knowledge graphs accumulate and convey knowledge of the real world. It has been well-recognized that knowledge graphs effectively represent complex information; hence, they rapidly gain the attention of academia and industry in recent years. Thus to develop a deeper understanding of knowledge graphs, this paper presents a systematic overview of this field. Specifically, we focus on the opportunities and challenges of knowledge graphs. We first review the opportunities of knowledge graphs in terms of two aspects: (1) AI systems built upon knowledge graphs; (2) potential application fields of knowledge graphs. Then, we thoroughly discuss severe technical challenges in this field, such as knowledge graph embeddings, knowledge acquisition, knowledge graph completion, knowledge fusion, and knowledge reasoning. We expect that this survey will shed new light on future research and the development of knowledge graphs.

Ontology-Based Linked Data to Support Decision-Making within Universities

Article

Full-text available

Sep 2022

In recent years, educational institutions have worked hard to automate their work using more trending technologies that prove the success in supporting decision-making processes. Most of the decisions in educational institutions rely on rating the academic research profiles of their staff. An enormous amount of scholarly data is produced continuously by online libraries that contain data about publications, citations, and research activities. This kind of data can change the accuracy of the academic decisions, if linked with the local data of universities. The linked data technique in this study is applied to generate a link between university semantic data and a scientific knowledge graph, to enrich the local data and improve academic decisions. As a proof of concept, a case study was conducted to allocate the best academic staff to teach a course regarding their profile, including research records. Further, the resulting data are available to be reused in the future for different purposes in the academic domain. Finally, we compared the results of this link with previous work, as evidence of the accuracy of leveraging this technology to improve decisions within universities.

Educational Resource Search in Scottish Schools

Conference Paper

Mar 2024

Vidminas Vizgirda

Dynamic Analysis of Engineering Structures Based on a Remote Laboratory

Conference Paper

Sep 2023

Review of Linked Data and Open Educational Resources in Academia: Potential Benefits and Challenges

Article

Full-text available

Feb 2024

OntoMathEdu Educational Ontology: Problems of Ontological Engineering

Article

Sep 2023

Knowledge Graphs: Opportunities and Challenges

Article

Full-text available

Apr 2023
ARTIF INTELL REV

Design of a trust system for e-commerce platforms based on quality dimensions for linked open datasets

Article

Full-text available

Jan 2023

This article describes a proposal about a trust system for e-commerce platform based on semantic web technologies and trust dimensions rules. We try to expose a system that allow to manage communication processes between e-commerce platforms and users in a trustworthy manner. It allows the data flows and transactions gain more trust across the entire process. All of this can be achieved through the inference of rules exposed in the defined ontology, complemented by a cloud-based system with microservices architecture. With the implementation of the system through an e-commerce platform, could consume data from the microservices in order to get inferences about its clients that want to buy or sell something within its system. This system was created based on rules defined by the ontology, as well as the microservices could be used to register information about multiple e-commerce transactions. The result of this work is the Ontology and semantic web rules defined and implemented through protege.

e-LION: Data integration semantic model to enhance predictive analytics in e-Learning

Article

Sep 2022
EXPERT SYST APPL

In the last years, Learning Management systems (LMSs) are acquiring great importance in online education, since they offer flexible integration platforms for organising a vast amount of learning resources, as well as for establishing effective communication channels between teachers and learners, at any direction. These online platforms are then attracting an increasing number of users that continuously access, download/upload resources and interact each other during their teaching/learning processes, which is even accelerating by the breakout of COVID-19. In this context, academic institutions are generating large volumes of learning-related data that can be analysed for supporting teachers in lesson, course or faculty degree planning, as well as administrations in university strategic level. However, managing such amount of data, usually coming from multiple heterogeneous sources and with attributes sometimes reflecting semantic inconsistencies, constitutes an emerging challenge, so they require common definition and integration schemes to easily fuse them, with the aim of efficiently feeding machine learning models. In this regard, semantic web technologies arise as a useful framework for the semantic integration of multi-source e-learning data, allowing the consolidation, linkage and advanced querying in a systematic way. With this motivation, the e-LION (e-Learning Integration ONtology) semantic model is proposed for the first time in this work to operate as data consolidation approach of different e-learning knowledge-bases, hence leading to enrich on-top analysis. For demonstration purposes, the proposed ontological model is populated with real-world private and public data sources from different LMSs referring university courses of the Software Engineering degree of the University of Malaga (Spain) and the Open University Learning. In this regard, a set of four case studies are worked for validation, which comprise advance semantic querying of data for feeding predictive modeling and time-series forecasting of students’ interactions according to their final grades, as well as the generation of SWRL reasoning rules for student’s behaviour classification. The results are promising and lead to the possible use of e-LION as ontological mediator scheme for the integration of new future semantic models in the domain of e-learning.

A rating system that open-data repositories must satisfy to be considered OER: Reusing open data resources in teaching

Conference Paper

Full-text available

Apr 2017

Guidelines to producing structured interoperable data from Open Access Repositories

Conference Paper

Full-text available

Oct 2016

One of the fundamental concepts of Open Educational Resources (OER) is “the ability to freely adapt and reuse existing pieces of knowledge.” The application of Semantic Web approach and Linked Data technologies to Open Education seeks to turn data and metadata from open educational repositories into actionable interoperability for the improvement of discovering, using and reusing of OER. Interoperability is not an end in itself. Instead, optimizing the level of interoperability has societal and educational value as a means to others purposes. Interoperability can have a positive impact on open innovation, user choice, ease to reuse and adapt educational materials, global discovery of open and diverse content, among other things. This paper reports on the implementation of Linked Open Data for open access repositories in a new interoperable and global open educational ecosystem. The goal is to improve the metadata interoperability between various collections of open material, so as to facilitate the discoverability and subsequent combining, remixing, or adapting OER; that is, OER data should be easily accessible to any user: human being or a machine agent. This work addressed two challenges in the OER ecosystem: providing evidence of globally discoverability and reusability academic resources. Although there is much further potential for teaching and learning to realize, linked open data is a critical enabler of global and interoperable OER ecosystem.

Promotion of self-learning by means of Open Educational Resources and semantic technologies

Conference Paper

Full-text available

Jun 2015

How Can Ontology Design Patterns Help Ontology Refinement?

Article

Full-text available

Jan 2014

Correlational Analysis Between School Performance and Municipal Indicators in Brazil Supported by Linked Open Data

Conference Paper

Full-text available

Apr 2016

Bruno Penteado

The advance in quality of public education is a challenge to public managers in contemporary society. In this sense, many studies point to the strong influence of socioeconomical factors in school performance but it is a challenge to select proper data to perform analyses on this matter. In tandem, it has happening a growth in provision of big quantities of educational indicators data, but in isolate cases, and by different agencies of Brazilian government. For this work, we use both education and economic indicators for analysis. The following socioeconomical indicators were selected: municipal human development index (MHDI), social vulnerability index (SVI), Gini coefficient and variables extracted from DBpedia, as part of the connection of this data to the Web of data: GDP per capita and municipal population. These data were used as independent variables to look into their correlations with Brazilian Basic Education Development Index (IDEB) performances at municipal level, supported by the application of linked open data principles. OpenRefine was used to extract the data from different sources, convert to RDF triples and then the mapping of the variables to existing ontologies and vocabularies in this domain, aiming at the reuse of existing semantics. The correlational analysis of the variables showed coherence with the literature about the theme, with significative magnitude between IDEB performances and the indicators related to income and parent education (SVI and HDI), besides moderate relations with the other varibles, except for the municipal population. Finally, the consolidated dataset, enriched by information extracted DBpedia was made available by a SPARQL endpoint for queries of humans and software agents, allowing other applications and researchers to explore the data from other platforms.

Semantic Annotation for Knowledge Management: Requirements and a Survey of the State of the Art

Article

Jan 2006

LDOW2017: 10th Workshop on Linked Data on the Web

Conference Paper

Apr 2017

The 10th Linked Data on the Web workshop (LDOW2017) was held in Perth, Western Australia on April 3, 2017, co-located with the 26th International World Wide Web Conference (WWW2017). In its 10th anniversary edition, the LDOW workshop aims to stimulate discussion and further research into the challenges of publishing, consuming, and integrating structured data on the Web as well as mining knowledge from said data.

Resource Description Framework (RDF) Model and Syntax Specification

Technical Report

Jan 1999

Analysing and Improving Embedded Markup of Learning Resources on the Web

Conference Paper

Apr 2017

Web-scale reuse and interoperability of learning resources have been major concerns for the technology-enhanced learning community. While work in this area traditionally focused on learning resource metadata, provided through learning resource repositories, the recent emergence of structured entity markup on the Web through standards such as RDFa and Microdata and initiatives such as schema.org, has provided new forms of entity-centric knowledge, which is so far under-investigated and hardly exploited. The Learning Resource Metadata Initiative (LRMI) provides a vocabulary for annotating learning resources through schema.org terms. Although recent studies have shown markup adoption by approximately 30% of all Web pages, understanding of the scope, distribution and quality of learning resources markup is limited. We provide the first public corpus of LRMI extracted from a representative Web crawl together with an analysis of LRMI adoption on the Web, with the goal to inform data consumers as well as future vocabulary refinements through a thorough understanding of the use as well as misuse of LRMI vocabulary terms. While errors and schema misuse are frequent, we also discuss a set of simple heuristics which significantly improve the accuracy of markup, a prerequisite for reusing learning resource metadata sourced from markup.

A comprehensive quality model for Linked Data

Article

Jan 2017

With the increasing amount of Linked Data published on the Web, the community has recognised the importance of the quality of such data and a number of initiatives have been undertaken to specify and evaluate Linked Data quality. However, these initiatives are characterised by a high diversity in terms of the quality aspects that they address and measure. This leads to difficulties in comparing and benchmarking evaluation results, as well as in selecting the right data source according to certain quality needs. This paper presents a quality model for Linked Data, which provides a unique terminology and reference for Linked Data quality specification and evaluation. The mentioned quality model specifies a set of quality characteristics and quality measures related to Linked Data, together with formulas for the calculation of measures. Furthermore, this paper also presents an extension of the W3C Data Quality Vocabulary that can be used to capture quality information specific to Linked Data, a Linked Data representation of the Linked Data quality model, and a use case in which the benefits of the quality model proposed in this paper are presented in a tool for Linked Data evaluation.

Linked Data in Education: A Survey and a Synthesis of Actual Research and Future Challenges

Abstract and Figures

Recommended publications

The Canadian health census as Linked Open Data: towards policy making in public health

Towards an Ontology for Conceptual Modeling

DYNAMIC SEMANTIC USER PROFILING FROM IMPLICIT WEB NAVIGATION DATA

Linked Data and the Charm of Weak Semantics: Introduction: The Strengths of Weak Semantics