Linked data for improving student experience in searching e-learning resources

(1)

2011

Linked data for improving

student experience in

searching e-learning resources

Master Thesis Project

Author:

Julieth Patricia Castelanos Ardila

Advisors:

Annabella Loconsole Marie Gustafsson Friberger

Examiner: Paul Davidsson

(2)

Linked data for improving student experience in searching e-learning resources

2

2011

Acronyms

BIBO BIBliographic Ontology - provides main concepts and properties for describing citations and bibliographic references (i.e. quotes, books, articles, etc) on the Semantic Web. - http://bibliontology.com/specification

CC Creative Commons - Nonprofit organization that develops, supports, and stewards legal and technical infrastructure that maximizes digital creativity, sharing, and innovation. - http://creativecommons.org/

DCMI Dublin Core Metadata Initiative - a non-profit organization engaged in the development of interoperable metadata standards. - http://dublincore.org/

DOAP Description Of A Project vocabulary - IT is a project to create an XML/RDF vocabulary to describe open source projects. - http://usefulinc.com/ns/doap

FOAF The Friend Of A Friend (FOAF) vocabulary - project devoted to linking people and information using the Web. - http://xmlns.com/foaf/spec/

HTML: HyperText Markup Language - the predominant markup language for web pages. -

HTTP: HyperText Transfer Protocol - a networking protocol for distributed, collaborative, hypermedia information systems. -

IEEE The Institute of Electrical and Electronics Engineers - http://www.ieee.org/index.html

LOD Linking Open Data cloud diagram - Datasets that have been published in linked data format, by contributors to the Linking Open Data community project and other individuals and organisations. -

LOM Learning Object Metadata - an IEEE standard for metadata descriptions of learning objects. -

MOAT Meaning Of A Tag - It defines a lightweight ontology to represent how different meanings can be related to a tag. - http://moat-project.org/ontology

(3)

3

2011 specifications. - http://www.openarchives.org/

OWL Web Ontology Language – It is a modelling language for expressing formal semantics of RDF properties and classes. -

RDF: Resource Description Framework - a W3C specification for metadata - descriptions. - http://www.w3.org/RDF/

RDFS The RDF Vocabulary Description Language, also known as RDF Schema

SIOC Semantically-Interlinked Online Communities - provides the main concepts and properties required to describe information from online communities (e.g., message boards, wikis, weblogs, etc.) on the Semantic Web -

http://www.w3.org/Submission/sioc-spec/

SKOS Simple Knowledge Organization System - a W3C specification for represent knowledge organization systems such as thesauri or taxonomies using RDF. - http://www.w3.org/2004/02/skos/

SPARQL: SPARQL Protocol and RDF Query Language - a W3C query language for RDF. - http://www.w3.org/TR/rdf-sparql-query/

URI: Universal Resource identifier - a globally unique identifier designed to be used on the WWW. -

URL Uniform Resource Locator – It is the unique address for a file that is accessible on the internet. -

WGS84 The basic RDF vocabulary that provides the Semantic Web community with a namespace for representing lat(itude), long(itude) and other information about spatially-located things - http://www.w3.org/2003/01/geo/

(4)

4

2011

5

2011

3.1.8. Use of vocabularies in the data sets in the LOD cloud diagram ... 36

3.1.9. Data sources that provide origin metadata ... 38

3.1.10. Data sources that provide licensing metadata ... 39

3.2. THE INTERNET IN EDUCATION, AND SUITABLE SOURCES OF e-LEARNING RESOURCES. ... 40

3.2.1. Web 2.0 and collaborative e-learning ... 40

3.2.2. Learning resources and their potential ... 41

3.2.3. Educational issues in front of technological mediations ... 42

3.3. SURVEYS RELATED TO SEARCHING E-LEARNING RESOURCES BY E-LEARNERS. ... 43

4. INVESTIGATION OF THE METHODS USED BY STUDENT FOR EXPLORING AND DISCOVERING e-LEARNING RESOURCES ... 45

4.1. DATA ANALYSIS AND INTEPRETATION ... 45

4.1.1. Data collection ... 45

4.1.2. Data analysis ... 45

4.1.3. Answers to the key research questions of the questionnaire ... 55

4.2. INTERPRETATION OF THE RESULTS ... 57

4.3. THREATS TO VALIDITY ... 58

5. THE PROTOTYPE DESIGN ... 60

5.1. FEATURES OF A e-LEARNING ENVIRONMENT PROTOTYPE... 60

5.1.1. Gaps found in the state-of-the art ... 60

5.1.2. A design proposal of an e-learning environment ... 62

5.1.3. The role of the teacher in the e-learning collaborative environment ... 64

5.1.4. Selection of the e-learning contents. ... 66

5.1.5. Comments and rankings ... 67

5.2. REQUIREMENTS SPECIFICATION ... 68

5.2.1. Functional requirements for the e-learning collaborative environment ... 69

5.3. QUALITY REQUIREMENTS FOR THE E-LEARNING COLLABORATIVE ENVIRONMENT ... 70

5.4. STAKEHOLDERS ... 70

5.5. USE CASES DIAGRAMS ... 71

5.5.1. General use case for student Interactions ... 71

5.5.2. Use case index ... 72

5.6. ARCHITECTURE OF THE e-LEARNING COLLABORATIVE ENVIRONMENTS WITH LINKED DATA TECHNIQUES ... 73

5.6.1. Transform mapping ... 74

(6)

6

2011 6. REQUIREMENTS VALIDATION ... 78 6.1. REQUIREMENTS TRACEABILITY ... 79 6.1.1. Traceability matrix ... 79

6.2. REQUIREMENTS REVIEW CHECKLIST ... 80

6.3. LIST OF PROBLEMS AND LIST OF ACTIONS ... 81

7. CONCLUSIONS ... 82

7.1. SUMMARY ... 82

7.2. CONTRIBUTION ... 83

7.2.1. Research justification ... 83

7.2.2. Answering the research questions ... 84

7.3. FUTURE WORK ... 85

8. REFERENCES ... 87

9. APPENDICES ... 93

9.1. APPENDIX A: SURVEY ABOUT e-LEARNING RESOURCES ... 93

9.2. APPENDIX B: MAPPING BETWEEN THE AVAILABILITY OF DATA SOURCES IN THE LINKED DATA COMMUNITY AND THE PREFERENCES OF THE STUDENTS ... 99

9.2.1. Contents selected by students and their availability in the LOD cloud diagram ... 99

9.2.2. According to Learning Object Metadata (LOM) criteria ... 105

9.2.3. Dataset selected and their vocabularies ... 105

9.2.4. How comments, rankings and citations of the e-learning resources could be addressed with the Linked data approach ... 106

9.2.5. Reliability in Contributors of e-learning resources, and their availability en in the LOD cloud diagram ... 106

9.3. APPENDIX C: USE CASES TEMPLATES ... 107

9.3.1. Provide Resources ... 107

9.3.2. Explore e-learning Resources ... 108

9.3.3. Use Resources ... 109 9.3.4. View Revisions... 110 9.3.5. Manage Account ... 111 9.3.6. Create Account ... 113 9.3.7. Delete Account ... 114 9.3.8. Modify Account ... 115

9.3.9. Manage Personal resources information ... 116

9.3.10. Save resources information ... 117

(7)

7

2011 9.3.12. Modify resources information ... 119 9.3.13. Make revisions ... 120 9.3.14. Publish new Resources ... 121

(8)

8

2011

List of Figures

Figure 1: An RDF Graph Describing Eric Miller. [12] ... 27

Figure 2: A schematic representation of the hierarchy of elements in the LOM data model ... 30

Figure 3: Linking Open Data cloud diagram, http://lod-cloud.net ... 34

Figure 4: The distribution of triples by domain. ... 36

Figure 5: The distribution of links by domain. ... 36

Figure 6: The distribution of most widely used vocabularies ... 38

Figure 8: Discipline Area ... 46

Figure 7: Educational Level ... 46

Figure 9: Sites/Sources of e-learning resources most visited ... 47

Figure 10: Use of social networks for learning ... 48

Figure 11: Initial approach in searching e-learning resources ... 48

Figure 12: Criteria in searching e-learning resources ... 49

Figure 13: Resources recommended by teachers or other students ... 50

Figure 14: Choosing of resources with comments ... 50

Figure 15: Rankings in e-learning resources ... 51

Figure 16: Citations on articles and books ... 51

Figure 17: Sense of reliability in recognized sources of learning resources ... 52

Figure 18: Preferences in use of learning resources ... 53

Figure 19: Ways for saving URL´s ... 54

Figure 20: Possibility of making comments ... 54

Figure 21: Possibility of making rankings ... 55

Figure 22: Interest in belonging to a e-learning/s-science group ... 55

Figure 23: Contents in the e-learning environment prototype ... 63

Figure 24: Level decision in front of the selection of e-learning resources ... 68

Figure 25: General use case for students interaction ... 72

Figure 26: Data flow diagram for the environment ... 75

Figure 27: General architecture or the e-learning collaborative environment ... 76

(9)

9

2011

List of Tables

Table 1: Overview of the amount of triples as well as the amount of RDF links per domain in LOD

cloud diagram, Bizer, C. & Jentzsch A. [29] ... 35

Table 2: Vocabularies widely used in LOD cloud diagram and links provided to the data sources ... 37

Table 3: Provide origin metadata by topical domain ... 39

Table 4: Licensing metadata ... 39

Table 5: Threat to Internal Validity ... 58

Table 6: Threat to External Validity ... 59

Table 7: Two Enactments of a Formative Assessment Cycle, Minstrell et al [46]. ... 64

Table 8: Use case index of the project ... 73

Table 9: Traceability matrix ... 80

Table 10: Requirements review checklist ... 81

Table 11: Lectures in the LOD cloud diagram ... 100

Table 12: Online Research papers in the LOD cloud diagram ... 100

Table 13: e-Books in the LOD cloud diagram ... 101

Table 14: Tutorials/Manuals in the LOD cloud community ... 101

Table 15: Interactive resources in the LOD cloud diagram ... 101

Table 16: Statistics and government sites in the LOD cloud diagram... 102

Table 17: Wikipedia and the LOD cloud diagram ... 103

Table 18: Google scholar and the LOD cloud diagram ... 103

Table 19: Google maps and the LOD cloud diagram ... 103

Table 20: YouTube and the LOD cloud diagram ... 103

Table 21: Universities and the LOD cloud diagram ... 104

Table 22: Data Sources selected and their vocabularies ... 105

Table 23: Comments in things in LOD cloud diagram ... 106

Table 24: Use case for providing Resources ... 108

Table 25: Use case for exploring e-Learning Resources ... 109

Table 26: Use case for Using e- resources ... 110

Table 27: Use cases for viewing revisions ... 111

Table 28: Use case for managing account ... 112

Table 29: Use case for managing account ... 113

Table 30: Use case for deleting account ... 114

Table 31: Use case for deleting account ... 115

Table 32: Use case for managing personal resources information ... 116

Table 33: Use case for saving resources information ... 117

Table 34: Use case for deleting resources information ... 118

Table 35: Use case for modifying resources information ... 119

Table 36: Use case for making reviews ... 120

Table 37: Use cases for publishing new resources ... 121

(10)

10

2011

1. INTRODUCTION

1.1. PROJECT IDEA

The collection and the use of data on the internet with e-learning purposes are tasks made by many people every day, because of their role as teachers or students. The web provides several data sources with relevant information that could be used in educational environments, but the information is widely distributed, or poorly structured. Also, resources on the web are diverse, sometimes with high quality, but sometimes not. These situations involve a difficult search of e-learning resources, and therefore a lot of time invested, because the search process – typing, reasoning, selecting, using resources, bookmarking, and so forth - is completely executed by humans, despite that some of them can be executed by computers. Heath presents in [1], the linked data techniques as option to tackle the issues related to publishing and to exploring data on the internet, because these techniques are used for exposing, sharing and connecting the Web of Data, using Universal Resource Identifier (URIs) and Resource Description Framework (RDF) . Berners-Lee [2] expose the basis of linked data techniques and highlight the differences between the two modes of web information: the web of hypertext, and the web of data. Both are constructed with documents on the web, but the web of data describes information – it includes the connections - with RDF language, and the URI identifies objects or concepts – pages, people, resources, and so on - Instead, the web of hypertext uses relationships anchors in hypertext documents written in HTML (Hypertext Markup Language) and the URI concept is just for location. The features of the web of hypertext makes its information difficult to be crawled by machines. linked data contributes with the growing of the Web of data by applying four basic rules: ―Use URIs as names for things; Use HTTP (Hypertext Transfer Protocol) URIs so that people can look up those names; When someone looks up a URI, provide useful information, using the standards such SPARQL (Query Language for RDF); Include links to other URIs, so that they can discover more things”. Many efforts have been carried out in the last years using linked data techniques, and a great number of data sets are available for being used. Datasets available with dereferencable1 URIs are exposed in [3]. In summary, linked data provides designed practices for organizing, and for discovering information using the processing power of computers. At the same time, the community of

1

Dereferencing is the act of retrieving a representations of a resource identified by a URI. For more information, see section 2.1.2

(11)

11

2011 linked data provides data sets that are already connected, and this information could be consumed by people anytime as resources with e-learning purposes

Additionally to practices of linked data, this master thesis will address the techniques used by students when searching e-learning resources, through a survey. The resources used by the students, as well as the sources preferred, will be compare with the current resources offered by the linked data community. Likewise, the strategies and techniques selected by the students will be taken into account, in order to establish the basic requirements of a e-learning collaborative environment prototype

1.2. AIMS AND OBJECTIVES

1.2.1. Main Goal

Improve the collaborative e-learning experience, through the design of an e-learning environment based on linked data techniques.

1.2.2. Objectives

 Identify linked data techniques as well as their uses in existing e-learning models.  Identify students preferences in regarding of searching e-learning resources.

 Design the architecture of a prototype, based on the findings of the previous objective.

1.3. MOTIVATION

Learning is an area where web-based technology has played an important role in recent years. The web technology supports educational tasks, such as extracting/publishing resources for teaching-learning purposes, real-time interaction among people involved in the educational process, and so forth. Since Web 2.0 appeared, students have found many ways of interaction with mates, and then, keeping in touch virtually, for example to share learning contents. Teachers as well as educational institutions like Universities, colleges, and so on, have found the web as the mean to interchange contents, points of view and many other activities with partners, scientists and their pupils. The web involves many resources for learning; not only web pages with information, but also knowledge communities, collaborative networks, and expert systems. Also, the idea of sharing one´s thoughts about things with others, - micro blogs, comments, rankings, photos, etc. -

(12)

12

2011 persuades more and more people every day, because of the easy spreading web-based systems. Consequently, the web provides not only several data sources with useful and relevant information with e-learning purposes, but also information that is not easy to retrieve, and therefore wasted data . Sometimes, the information is inappropriate, and they must be filtered, in order to be relevant. Improving the way to exploit the web is an important step in the development of e-learning technology. The web would be a useful mechanism in the learning process if we take advantage of the information which is placed there. These improvements are related to the distribution of tasks: computers can be faster than humans in searching information with organized data. For humans, the search process on the web becomes a difficult task, and also very tedious, as a result of the large amount of information disposable to be consulted. For computers, large amounts of data is not a problem. Computers can be very useful, if the data is reassigned in standard formats, in order to make connections between them. Consequently, the computer can find the information easily using adequate search engines, that crawl the Web of Data by following links between data sources and provide expressive query capabilities over aggregated data, similar to how a local database is queried today, as Bizer et al. [4] suggests. According to this condition, the current web is not ready for the new age of computers reasoning, because data is published in several formats and the computers are unable to establish ways of connection without the human intervention.

The linked data approach is a semantic web practice that allows structure on the web of data using RDF triples. RDF is a standard model for data interchange on the Web and it is presented as a language for representing information about resources in the World Wide Web [5]. Data can become structured information with RDF. This process is called Rdfization. The structured information makes easier processing on computers, not only for using the information, but also for creating better connections around the world wide web. e-Learning resources could be reached, if the web of data enables the connections between several sources on the internet. Universities, scientific sites as well as journals, have published their own information for free, and these sources could be seriously considered when we search e-learning resources. Metadata about the general resources provides information to e-learners in order to know about the provenance of the data. Consequently, the e-learner obtains the freedom to select the most convenient according with their educational needs. Taking into account the amount of data located on the internet and the opportunity to make connections between information and sources that provide usable information for e-learning purposes, and also, the datasets currently connected in the linked data community, we could make the web a more interesting place, and also a relevant tool for e-learners, in order to improve their experience in searching e-learning resources.

(13)

13

2011

1.4. RESEARCH QUESTIONS

linked data is a technique that allows publishing and exploring data on the internet [1]. Searching information on the Web of Data is a process that is made by following links between data sources. This process provides expressive query capabilities over aggregated data, similar to how a local database is queried today [4]. This technique can be useful for crawling data on the internet for educational purposes, because learning is an area where web-based technology has played an important role in recent years, and lot of instructional information have been placed around the World Wide Web, using Learning Object Metadata (LOM). LOM has established a standard (base schema) which defines a structure for interoperable descriptions of learning objects [17]. In this context, our main research question is:

RQ1: How could linked data support the collection of information on the internet in order to enrich collaborative e-learning environments?

The research must combine a depth literature review, in order to understand the current state of the art about the techniques around linked data issues and metadata in learning objects as well as the current preferences of the students in searching resources for self learning. The two following questions are placed for helping in answering these important aspects.

RQ2: What are the remarkable features offered by existing learning environments based on linked data techniques?

RQ3: To what extent are the students’ preferences in searching e-learning resources supported by the current search engines and/or learning environments?

1.5. EXPECTED OUTCOMES

By developing this project, we are expecting to gain knowledge about searching e-learning resources, which use linked data. We are interested in establish connections between theoretical assumptions about linked data Techniques, and current practices developed by students when they search resources on the internet, in order to conduct future research around this topic. The outcomes of this project will support the answering of the research questions following the order of the thesis:

(14)

14

2011 a. Linked data techniques and its uses in e-learning resources exploration

b. The internet in education, and suitable sources of e-learning resources. c. Surveys related to searching e-learning resources by e-learners

RQ3 will be answered with the results of a survey about the current techniques for searching e- learning resources on the internet, used by the students in Malmö University

RQ1: will be answered by studying the following aspects:

a. Mapping between the availability of datasets in the Linking Open Data Cloud diagram and the current preferences of the students:

 Data sets selected from the LOD (Linking Open Data) cloud diagram: The results of the survey will be contrasted with the information published using linked data techniques, in order to establish the availability of data sources for e-learning purposes.

 Identification of vocabularies used in the linked data community: Discovering and understanding the vocabularies used in the data sets selected for e-learning purposes

b. Design of an e-learning collaborative environment prototype, using the findings of the research

 Prototype overview: General goals of the prototype design.

 Requirements specification: The basic requirements derived from the survey will be integrated in a model using UML use cases

 General design of the e-learning collaborative environment: which allows the connection between students preferences found in the requirements elicitation, with the current data available on the internet, and the new ones published by the users.

1.6. OUTLINE

Chapter 2 discusses the research methodology as well as the constructing and administering the research survey in which the current thesis based the requirements elicitation.

Chapter 3 lays the groundwork for the rest of the thesis by presenting the principles and terminology of linked data, as well as related work about the internet in education, availability

(15)

15

2011 of e-learning resources, and surveys about the use of the internet in e-learning resources searching

Chapter 4 presents the investigation of the methods used by student for exploring and discovering e-learning resources through the data analysis and interpretation of the survey.

Chapter 5 Introduces to the prototype design. It includes the prototype idea, the requirements specification using the data analysis of the survey, and the architecture of the e-learning collaborative environment using the assumptions reached in the literature review and the dereferencable URIs found in the linked open data cloud diagram. The design of components in the environment will be addressed in terms of UML diagrams.

Chapter 6 Validates the requirements of the prototype.

Chapter 7 Tackles conclusions of the master thesis project in order to find incomes for further research in the area. This chapter also shows the contributions e-learning world evaluation is based on the benefits indentified by using this approach and gives indications of what future work can be done to improve the results.

(16)

16

2011

2. THE RESEARCH METHODOLOGY

In this section it is to describe the methodological research assumptions underpinning the present work. It includes the general research strategy as well as the data collection method. Furthermore, this section defines boundaries of the research design, and situates the research amongst existing research traditions in computer science.

2.1. METHODOLOGY OF THE THESIS PROJECT

2.1.1. Overview about the research methodologies

The mixed-methods research is the methodological approach selected to address the present work. This approach was selected, taking into account the nature of the research in which take part computer science techniques and educational research. The assumptions of Gall [6] validate the selection of the method because he explains that this is a way of have broader understanding of the problem, making richer analysis: ―A review of quantitative studies about a particular phenomenon combined with a review of qualitative studies about the same phenomenon can provide richer insights and raise more interesting questions for future research than is only one set of studies is considered‖. This kind of methodology includes an extensive data collection and analysis of both kind of data: numeric and text.

The planning procedure in mixed methods is important and requires special attention. For this reason, it is useful to take into account the aspects that influence the designing of the method study as Creswell [7] explains ―Four important aspects are: timing, weighting, mixing and theorizing‖. In the case of the current research, these aspects will be addressed as follows:

 The timing of the qualitative and quantitative data collection is concurrently, that means: the information is gathered at the same time and the implementation is simultaneous.  The weighting means priority given to the approaches selected. In this case, the

importance of the deductive approach allows to emphasize in the quantitative information to express, in better way, the hypothesis raised.

 The Mixing the data occurs in the three stages of the research: data collection, the data analysis and interpretation. The mixing consists of integrating the numbers with the text provided by questionnaire respondents.

(17)

17

2011  The Theorizing is tackling the theoretical perspective established in the literature review

and the understanding of the needs of the subjects in regards to search e-learning resources. It will be used an implicit theorization, that means theories are not discussed in front of the results of the survey because they will be already embedded in the creation of the survey and its subsequent analysis.

In essence, the strategy selected for addressing this research is a concurrent transformative strategy because as Creswell [7] explains "it is guided by the researcher´s use of a specific theoretical perspective as well as the concurrent collection of both quantitative and qualitative data". The concurrent transformative model is applied in data collected at the same time, prioritizing the quantitative one over the qualitative, because, qualitative data are used to explain some further issues of the questions in the survey in order to enrich the answers given by the respondents. This kind of approach is called Explanatory Sequential.

The data collection method arises by this research is a survey. Surveys are used in research because allow to collect data about any topic, because surveys allow to collect data about Phenomena that are not directly observable: inner experience, opinions, values, interests and the likes, Gail [6].

The data collection tool supported by this research is a questionnaire managed by the respondent itself. This data collection method ensures reliability in the topic cited, because it allows respondents freedom for writing their own answers without the intervention of a interviewer, as well as anonymity. Despite the current research is based on mixed research approach, interviews are not conducted to gather any data. The questionnaire is designed to permit textual answers by the respondents after every close-ended question, in order to strengthen them with their particular comments or opinions.

2.1.2. Scientific Hypotheses

If we develop a methodology for searching e-learning resources based on linked data approach, we develop e-learning environments more structured and more relevant, in order reduce the time for exploring, selecting, and use e-learning objects, as well as a framework for selecting reliable resources

(18)

18

2011

2.1.3. Traditional Research in Computer Science

The present project implies the design of a methodology for improving the students experience in searching e-learning resources in order to set the foundations for new developments in the field of e-learning involving semantic web techniques. The questionnaire and its analysis will work as a elicitation of requirements in the normal process of software development. To compare the requirements elicited with the state-of-art are the second step in this research in order to configure a path for making a design that really contributes with the purpose of the project. Then, the process will cover data design as well as internal linkages proposal between data sources provided with the linked data community and new ones considered relevant in the present research in order to follow the practices involved in the linked data approach.

2.1.4. Data Collection Method

The assumption underlying in the survey is that the students, in general, have adopted the information offered by the Web as a basic resource to consult their informational shortcomings, and used this data in their learning experience. The research will be used both descriptive and exploratory approach in order to understand the solutions of the research questions. At the beginning, the descriptive research will be used to get data about the current status of the phenomenon through a survey. Then, the exploratory research will be used to yield information discovered by other researchers. This assumptions allow the researcher familiarize herself with the concepts of the problem under study to facilitate development of insights. The present study is an exploratory attempt since it would try to gather information regarding the behaviour that the students in Malmö University show in front of exploration, selection and using of e-learning resources . The researcher makes use of existing literature in order to verify the information gathered with the surveys and come up with preliminary ideas regarding the research problem.

The questionnaire will use close and open-ended questions as well as questions using Likert Scale answers. The instrument is at the respondent hands over a paper-based questionnaire. There are no participation of any interviewer in the procedure of the questionnaire answers, just at the beginning, in the introduction of the survey.

(19)

19

2011

2.2. CONSTRUCTING AND ADMINISTERING THE RESEARCH

QUESTIONNAIRE

A relevant questionnaire must be structured in a precise way. In order to achieve relevant data, the procedure to follow will be obtained from some of the steps proposed by Neuman [8]: (1) defining survey objectives, (2) selecting a sample, (3) designing the questionnaire format, (4) pretesting the questionnaire, (5) precontacting the sample, (6) writing a cover letter and distributing the questionnaire, (7) following up with nonrespondents, and (8) analyzing questionnaire data. This approach is interesting and has a valuable framework for understand the whole process that must be performed in the current research.

2.2.1. Defining Questionnaire Objectives

People involved in this data collection are students from Malmö University that are carrying out different careers, not only as undergraduate or postgraduate students, but also people who take single courses as well as college students.

At the moment of this research, we are interested in making a broad descriptive study, neither to specify nor to compare different subgroups. The aspect of the topic to study is related to the methods that students use for exploring and discovering e-learning resources and also, the opinions they have in front of these resources, specially reliability, usability and usefulness. Finally, facts and attitudes gathered from the questionnaire will be used for interpret the information, related to preferences of students from Malmö University in front of uses of e-learning resources and develop a theory from the findings.

 Questionnaire Aim

Investigate the methods used by students for exploring and discovering e-learning resources as well as the opinion they have about these resources in terms of reliability, usability and usefulness.

 Questionnaire Objectives

The survey objectives are related to the research questions and the survey would seek answers to this question. The objectives are broadly spelt out as follows:

a. To recognize the demographics of the population in terms of education level and discipline area studied in Malmö University.

(20)

20

2011 b. To investigate the types of exploration or search used by students in their search for

e-learning resources

c. To find out if the students take into account other people´s suggestions when choosing of their e-learning resources

d. To ascertain the reliability feeling in the sources of e-learning resources used by students

e. To find out the preferences that students have in the use and manage or their e-learning resources

f. To find out how the student evaluates the e-learning resources supplied by the internet as well as the sharing intention with people that have similar interest

2.2.2. Designing the questionnaire format

The questionnaire created are designed to request large amount of information around the exploration and the uses of e-learning resources, because the nature of the research was based on several LOM items (Learning Object Metadata). However, keeping the questionnaire as short as possible was a remarkable issue to solve. The first task consisted in organize the items, in order to confer logical sequence answers in the respondents, using key research questions. Then, questions were opened, in order to gain deeply understanding on every main question.

 Key Research Questions of the Questionnaire

There are 6 questions closely linked with the objectives of the survey. The questions are as following:

Question 1: What is the educational level and discipline that respondent is studying in Malmö University?

Question 2: How do the students explore the internet for searching e- learning objects?

Question 3: How important are suggestions of other people for the selection of e-learning resources?

Question 4: What sources of e-learning resources shows more reliability in the student´s preference?

Question 5: What kind of e-learning resources the students prefer and how they manage the source of that resource?

(21)

21

2011 Question 6: How do the student would assess the usefulness of the online learning resources as well as their intentions in sharing e-learning resources with mates?

 Splitting the main questions

This section was created to explain how the main questions were forked , in order to create a more accurate questionnaire. The main questions are enumerated and subquestions are presented below these numberings

Question 1

 Which educational level describes the respondents?  What discipline area the respondents are pursuing?

Question 2

 What kind of pages do the students usually visit for searching online learning objects?

 Do the students use social networks to find e-learning resources?

 What is the initial approach of the students when they search for e-learning resources?. How the students combine words to make a more precise search of their e-learning resources?

 Are the students interested in search their learning resources using criteria like author´s name, field, topic, date, language, format, learning resource type, interactivity level or context?

Question 3

 Are the students interested in select online learning resources recommended for teachers or other students?

 Are the students interested in select online learning resources commented by somebody else?

 Are the students interested in the rankings people do about the usefulness e-learning resources?

 Are the people interested in the amount of citations that the articles have?

Question 4

 Do the students trust in sources of e-learning content like YouTube, Wikipedia, Google and so forth?

(22)

22

2011 Question 5

 What kind of e-learning resources the students prefer to use?  How do the students remember the relevant web page URL´s?

Question 6

 Are the students willing to make comments about the quality of the learning object found?

 Are the students interested in rank the usefulness of the e-learning resource they have used?

 Are students interested in use social networking collaborative learning?

 Survey Instrument Completed

The survey instrument resulting in a questionnaire with 16 questions. It is able to tackle demographic information as well as the objectives given before. Finally, six sections are involved in the questionnaire.

 The first section (questions 1 and 2) of the questionnaire gathered the respondents’ demographic characteristics such as their education level and discipline area of study.

 The second section of the questionnaire (questions 3, 4, 5 y 6) collected the type of exploring respondents use in their e-learning resources search.

 The Section 3 of the questionnaire (questions 7, 8, 9 and 10) highlighted the importance of suggestions of other people for the selection of e-learning resources.

 The Section 4 of the questionnaire (question 11) was designed as a Likert scale to find out the reliability students have in some sources of e-learning resources .  The fifth section of the questionnaire (questions 12 and 13) focused on

preferences of the students for use resources available and how they manage the Url´s of this resources.

 The last section of the questionnaire (questions 14, 15 y 16) was specially included to find out the students orientation about assessing and sharing e-learning resources.

(23)

23

2011

2.2.3. Pilot testing the questionnaire

The survey questionnaire was first piloted to a sample of 11 students of Malmö University on 23 February 2011, during a class in which the participants have different backgrounds, level of study as well as different study areas. The comments and the answers from the respondents were taken into consideration during the process of refining the questionnaire. The primary reason for this is to develop questions that are relevant and which could be understood easily by the respondents. The questions were further thoroughly checked for reliability and validity. The final product is a 4-pages questionnaire that was used for the survey. The questionnaire is shown in Appendix A.

2.2.4. Target population and sample

The specific pool of population that the current research wants to study are the students of Malmö University which consists of 25 000 students enrolled in full- or part-time studies 20102. For tackling the current research, it was selected Nonprobability Sampling, which Neuman [8] explains as follow: "This means researchers rarely determine the sample size in advance and have limited knowledge about the larger group or population from which the sample is taken" But, according, also with Neuman [8], the specific type of Nonprobability sample selected was Quota Sampling, because the research designed is sure about the type of categories of students to investigate. These categories are related to the educational level that the students of Malmö University are pursuing: College, Undergraduate, Master, Doctoral and single courses.

In order to improve the quota sampling, the individuals that answered the questionnaire were selected in several places of the university, such as Orkanen Library, Kranen Library, and rooms in which the teacher allowed the data collection in class. Other questionnaires was collected in Celcuisgården accommodation, which have many students who belong to Malmö University. This procedure ensures the correct representation of all categories designed. The final data collection was of 82 questionnaires that represents 0,328 percent of the total population in these educational level categories in Malmö University

(24)

24

2011

2.2.5. Precontacting the sample

The rate of response is always a great problem to solve. In many cases, people are not interested in answering surveys, because they need the time for many other personal issues. Neuman [8] suggests that some kind of precontact, increases the rate of response. In the current survey the precontact was made in verbal way, opposite to the respondents and before give them the questionnaires, explaining the nature of the research as well as the importance of the reliable answers. Also, the teachers who allowed their classes to pick respondents up, offered time to think and answer the questions.

2.2.6. Writing a cover letter

A cover letter is used to explain the purpose of the study concerned. It needs to be brief, but it must convey certain information and impression as Neuman [8] explains: "to persuade the respondents that the study is significant and their answers are important". For purposes of this study, a cover letter was not created, because the preliminary information was given to the respondents verbally. However, a short presentation was written at the beginning of the questionnaire.

2.2.7. Following up with nonrespondents

As the surveys was doing in classrooms, libraries and in the accommodation corridor, the respondents were encouraged to filling the questionnaire at the same time it was given to them. Nevertheless, some surveys were not returned in the same moment of the application. These unanswered questionnaires were channelled in the next class, or in next meeting with the respondent, inquiring about them in a verbal way. 91, 11% of the surveys was collected.

(25)

25

2011

3. LITERATURE REVIEW

To give an introduction to linked data technology, this chapter starts by explaining the basic elements of the linked data approach, as well as the uses of this technique in exploration/publication of e-learning resources in Section 2.1. Following, in Section 2.2. we explain the influence of the internet in education and the availability of the resources with e-learning purposes. And finally, in Section 2.3 we describe the existing surveys related to searching e-learning resources by e-learners in different contexts.

3.1. LINKED DATA TECHNIQUES AND ITS USES IN E-LEARNING

RESOURCES EXPLORATION

3.1.1. Linked data overview

Computers today can show information for almost any source, but they have difficulties to crawl most of the information published, because the conventional data format i.e. HTML, is not sufficiently expressive to enable individual entities described in a particular document to be connected by typed links to related entities, Bizer et al. [4]. Computers have been considered as passive counters of information, because sometimes they just show the information that the human find, but their potential on searching data on the internet is wasted.

Semantic Web could make the computers a more active tools in order to help us in tasks like searching information, because semantic web refers to possibility of allowing the formal communication of a higher level of language (this is called metalanguage) between computers. Linked data is a Web semantic technique and refers to a set of best practices for publishing and interlinking structured data on the Web, Heath et al. [6]. Linked data conducts the possibility to configure a new kind of web called Web of Data. Heath et al. [9] as well as Bizer et al. [4], the principles of linked data:

1. URI references are not just Web documents and digital content, but also real world objects and abstract concepts;

(26)

26

2011 3. The Resource Description Framework (RDF) as Klyne et al. says [10] ―is a model that

has been designed for use in the context of the Web‖. RDF is used for publishing structured data on the internet and tools was created to crawl this data in a proper manner.

4. Finally, linked data context have types which describe the relationship between the things. This feature allows the discovery of more things.

3.1.2. Dereferencing URIs and Resource Description Framework (RDF)

Intelligent agents (IAs) “are software programs intended to perform tasks more efficiently and with less human intervention” Kravari et al [11]. It implies that this kind of complex software shall operate in the world wide web, performing a variety of tasks from caching and routing to searching, categorizing, and filtering. They may retrieve representations of resources by dereferencing URIs. Dereferencing is a useful concept in the context of semantic web, because, is described as the act of retrieving a representation of a resource identified by a URI, Lewis [12] Heath et al. [9], also implies ―Any HTTP URI should be dereferenceable, meaning that HTTP clients can look up the URI using the HTTP protocol and retrieve a description of the resource that is identified by the URI”

RDF allows URLs (Uniform Resource Locator) dereferencing because the data format could be determined by content negotiation inside the graph created. RDF was created for representing metadata about Web resources as well as for representing information about things that can be identified on the Web, even when they cannot be directly retrieved on the Web, Manola et al. [13]. For this reason, RDF is useful when information needs to be processed by applications, rather than HTML, which is oriented to display the information to people. RDF is based on the idea of identifying things using Web identifiers or URIs, and describing resources in terms of simple properties and property values. For this reason, RDF serves as mechanism for describing things.

RDF statement is composed by three basic parts: subject, predicate and object; Groups of statements are represented by corresponding groups of nodes and arcs. In Figure 1 there are a group of statements that refers to Eric Miller. This representation can be explained in words:

"there is a Person identified by http://www.w3.org/People/EM/contact#me, whose name is Eric Miller, whose email address is em@w3.org, and whose title is Dr."

(27)

27

2011

Figure 1: An RDF Graph Describing Eric Miller. [12]

3.1.3. Taxonomies, vocabularies and ontologies to describe data

RDF is an abstract data model for describing resources. RDF does not have any domain terms for describing classes, things, and relationships. For this reason, taxonomies, vocabularies and ontologies are designed and expressed in SKOS (Simple Knowledge Organization System), RDFS (the RDF Vocabulary Description Language, also known as RDF Schema) and OWL (the Web Ontology Language)

SKOS is described as a common data model for sharing and linking knowledge organization systems via the Web, Miles et al. [14]. The idea consists in sharing similar structure, and similar applications. SKOS is used to represent thesauri, taxonomies, subject heading systems, and topical hierarchies.

OWL is designed for being used by applications that need to process the content of information instead of just presenting information to humans, McGuinness et al. [15]. This Ontology allows that machines can read and interpret the web content created with XML,

(28)

28

2011 RDF, and RDF Schema (RDF-S) by providing additional vocabulary along with a formal semantics. RDFS and OWL are used in cases where subsumption relationships between terms, as well as inheritance, should be represented.

There are vocabularies created with different purposes. The recommendation is to reuse them instead of creating a different one, in order make the data being consumables by applications that may be tuned to well-known vocabularies. The following list, developed by Heath et al, [9], describes some vocabularies, used in common applications:

 The Dublin Core Metadata Initiative (DCMI) Metadata Terms vocabulary defines general metadata attributes such as title, creator, date and subject.

 The Friend-of-a-Friend (FOAF) vocabularydefines terms for describing persons, their activities and their relations to other people and objects.

 The Semantically-Interlinked Online Communities (SIOC) vocabulary (pronounced "shock") is designed for describing aspects of online community sites, such as users, posts and forums.

 The Description of a Project (DOAP) vocabulary (pronounced "dope") defines terms for describing software projects, particularly those that are Open Source.

 The Music Ontology defines terms for describing various aspects related to music, such as artists, albums, tracks, performances and arrangements.

 The Programmes Ontology defines terms for describing programmes such as TV and radio broadcasts.

 The Good Relations Ontology defines terms for describing products, services and other aspects relevant to e-commerce applications.

 The Creative Commons (CC) schema defines terms for describing copyright licenses in RDF.

 The Bibliographic Ontology (BIBO)provides concepts and properties for describing citations and bibliographic references (i.e., quotes, books, articles, etc.).

 The OAI Object Reuse and Exchange vocabulary is used by various library and publication data sources to represent resource aggregations such as different editions of a document or its internal structure.

 The Review Vocabulary provides a vocabulary for representing reviews and ratings, as are often applied to products and services.

 The Basic Geo (WGS84) vocabulary defines terms such as lat and long for describing geographically-located things.

(29)

29

2011 These vocabularies are designed to classify things - concepts, people, animals, etc - that can be used in a determined applications, create relationships, and define constraints on using those terms. In practice, vocabularies are to help data integration, and for this reason, they must be work together. This integration could be very complex to understand if they have several terms. Fortunately, there are metadata harmonization initiatives that provide frameworks to set up the vocabularies coordination in order to raising the expectations for metadata interoperability, Nilsson [16]. The relevance of this harmonization is defined in a successful exploration of the web. The metadata standards discussed in this thesis have all been chosen based on some kind of relevance for the field of e-learning and learning objects. The next section will be oriented completely to LOM (learning Object Metadata) because LOM is regarded as the dominant standard in this field.

3.1.4. Learning Object Metadata (LOM)

IEEE (The Institute of Electrical and Electronics Engineers) has established a standard (base schema) which defines a structure for interoperable descriptions of learning objects. It was called LOM (Learning Object Metadata). Learning Object in terms of the IEEE document is: "any entity -digital or non-digital- that may be used for learning, education or training‖ [17].

LOM metadata describes relevant characteristics of the learning object, and this is appropriate because the standard is suitable for facilitating search, evaluation, acquisition, and use of learning objects. LOMv1.0 base schema is composed by metadata in terms of a hierarchy of 76 elements classified into nine categories, specifying vocabularies, and allowing syntaxes for the value of each element. LOM metadata could be used to know not only metadata about for resources, but also information such as aspects of the lifecycle of a learning object, its pedagogical features, technical aspects and so on. Figure 2 shows a schematic representation of the hierarchy of elements in the LOM data model. Dublin Core Metadata Initiative has developed data elements related to the current LOM version. There are several works that talk about LOM standards, its usefulness and its features developed in learning environments. For instance, Svensson et al. [18] says that the IEEE-LOM1 metadata specification is a popular standard for enriching learning content with metadata to promote their reusability, discoverability and interoperability. This is a good starting point in terms of creation of vocabularies or reuse of existing vocabularies.

(30)

2011

Figure 2: A schematic representation of the hierarchy of elements in the LOM data model3

3

(31)

2011 Another citation of LOM are found in Santally et al. “The aim of those entities (Learning Objects) is to provide a tremendous set of learning knowledge that once developed can be exchanged among organisations, and be used to build individual lessons and courses”, [19].

LOM provides not only flexibility, but also a standardised description about the kind of objects used in education. Recently, Mogharreban et al. [20] realize that the conventional LOM lacks of control in their structure. Some of the problems are derived in impediment to reusability. They explain that Learning objects have been metaphorically described as pieces of something, like chunks, nugget s, LEGO™ blocks, Lincoln Logs™, atoms, molecular compounds, and crystals, that could be difficult to sort. For this reason, their proposal are related to a new vision of LOM, that encloses the use of a learning pod that "compass the continuum from discrete digital components to full-fledged aggregate coursework". With this approach, they have adopted a strictly granular approach to retain the highest level of reuse.

The use of LO (Learning Object) in the current research work is relevant. The general concept has been developed by many authors in order to lay out the principles of its foundation. Polsani [21], walked towards a clear concept of the reusable learning object. He says that any digital object can acquire the status of a learning object if it is wrapped in a learning intention, which has two aspects: form and relation. Form means the setting, context and environment for viewing the learning object, while relation refers interface that establishes a connection between the user and the bits of information stored in the computer memory banks. In the other hand, reusability suggests that the learning object is predisposed for reuse by multiple developers in various instructional contexts.

We are interested in understand the nine categories described in [17] and showed in Figure 2, and how they are applicable in the context of searching e-learning resources by students in Malmö University, selecting relevant features of each category and using them in the gathering of data from students. The complete standard is large and must be read in [17]

3.1.5. Browsing the Web of Data

The web of data has the ability to discover new things through the tracking of data. There are linked data browsers that allow users to navigate between data sources by following RDF links. Some of these browsers are the following:

(32)

32

2011  Disco hyperdata browser. Bizer et al, explains that it is a simple browser for navigating

the Semantic Web as an unbound set of data sources, [29] .

 The Tabulator browser. It is an user-friendly Semantic Web browser, Berners Lee et al, [23], that becomes a generic browser for linked data on the web. Heath et al, [9] also explains: ―the results of the query form a table that can then be analyzed with various conventional data presentation methods, such as faceted browsing, maps, timelines, and so on”

Any browser selected works with SPARQL. This is the query language used for querying RDF data. Its name is a recursive acronym that stands for SPARQL Protocol and RDF Query Language. The language works with a similar structure of SLQ language. SPARQL contains capabilities for querying required and optional graph patterns along with their conjunctions and disjunctions, Prud'hommeaux et al, [24].

3.1.6. e-Learning approaches for the linked data age

The web of linked data is a repository based on semantic technologies. Several researchers have been oriented to this kind of interoperable e-Learning repositories and establish that the Linked Data approach has the potential to fulfil the e-Learning vision of Web-scale interoperability of eLearning resources as well as highly personalised and adaptive eLearning applications. Dbpedia (http://dbpedia.org/) is the core of this approach and presents a first view related to data connection, using linked data techniques. Dbpedia as a community effort to extract structured information from Wikipedia, Auer et al. [25]. This project has mechanisms in which new authors are allowed to contribute in order to facilitate contents. Dbpedia dataset is accessible using three access mechanisms: linked data, the SPARQL protocol, and downloadable RDF. Also, It has been used in several e-learning projects, because their linked open datasets are suitable to be selected using the linked data methodology to almost any topic. Dbpedia project is still growing and it has open data A work based on weaving social e-learning platforms is the one described by Selver et al. [26]. They highlights the relationships among several vocabularies, such as FOAF (Friend of a friend), SIOC(Semantically-Interlinked Online Communities), and MOAT(Meaning of a tag) for interlinking data. These ideas present how is possible to use semantic web in order to enrich contents, using social networks and dbpedia. It explains an interlinking model for finding information, based on semantic enhancement of user contributed content in social e-learning platforms, using information about users, tags, and related resources. It includes the explanation of weaving process using the vocabularies: MOAT describes the meaning of

(33)

33

2011 tags, FOAF links to user profile and SIOC shows the representation of a user’s blog post to the FOAF profile. The ELGG community (http://community.elgg.org/), which is an open source social networking engine, provides the social e-learning platform as a start point for search the information related. The idea proposed by [26] is very useful, because the authors show the possibility to establish dialogue between several ontologies, thereby setting interoperability and enhance the navigability inside the ELGG community. Ways for combining RDF vocabularies is also tackled by Aleman et al. [27]. Contents source are subordinated to the production inside the community and dbpedia.

Video search with educational content was tackled by Waitelonis et al. [28]. This work establishes "the use Linked Open Data to complement already existing information about entities within the scope of the domain of our academic video search engine". This works shows how Yovisto's (http://www.yovisto.com) database provides recordings of speakers about scientific subjects. Yovisto, also has information about the source of this recordings such as name of the lecturer, country, city, type (university or other), and website. Dbpedia provides information about the university that is the source of the recording . The dbpedia dataset is integrate automatically inside Yovisto. Also, dbpedia is used for setting information about the speaker in the lecture recorded. The confidence in data is subject to DBPEDIA dataset only and it represents one of the major gaps of this work, because the reliability must be searched in more reliable sources, such as the universities that provides the recording. However, the approach achieved with this work is a fundamental contribution to the web of linked data.

In order to recognize the advances inside the linked data Community, and their possible uses in e-learning collaborative networking, we considered relevant to explain, briefly, the topology of the Web of data. Data sets are classified into the following topical domains: Media, Geographic, Publications, user-generated content, Government, cross-domain and Life sciences. Figure 3 shows the Linking Open Data cloud diagram, organized by topical domain, which are represented in different colors.

The organization of these domains must be understand for establishing a complete awareness of the possibilities that the Web of data has today, and how we can contribute for the creation of new data sets, that will implies the use of LOM. Our interest will focus in the publications and cross-domain, because these are suitable to LOM standard.

(34)

2011

Linked data for improving student experience in searching e-learning resources