Semantic Technologies: Linked Data and OER Opening and linking the data and content silos, to leverage the knowledge capital represented by our OER repositories Edmundo Tovar (UPM [email protected] ) Nelson Piedra (UTPL, [email protected]) | Jorge López UTPL, Janneth Chicaiza UTPL, Oscar Martínez UMH 2012 World Open Educational Resources Congress Wednesday 20 – Friday 22 June, 2012 Room XII, UNESCO HQ, Paris, France

#WorldOER #OpenEducationalResources #OpenCourseWare #linkeddata #ocw #oer #SemWeb #SemanticWeb #LOCWD #LOD this work is licensed under a Creative Commons Atribución-NoComercial-SinDerivadas 3.0 Ecuador License http://creativecommons.org/licenses/by-nc-nd/3.0/ec/

I. OER/OCW worldwide repositories a “pot a gold” In OER and OCW scope: Open license is Not Enough! A challenge to those involved in providing OER/OCW is establish ways where they can be most esasily found to use, reuse, sharing and remix. Our OER semantic vision: Educational Content + Open Licenses + Data in Machine Format

2

WHY OCW/OER + LINKED DATA?

In order to move forward and realize the promise of Linked Data for OCW/OER Repositories, Universities The Linked Data aid the discovery, reliable re-use of data, provide improved provenance and facilitate automated processing by increased flexibility to changes in presentation, use, reuse, remix and reduced ambiguity.

Linked Data is a question of…

•Open access for course, resources and materials •Legal compatibility of distributed educational resources silos •Improvement interoperability and accessibility of educational content •Best practices: • For identifiers (http and uris) • For modelling data (RDF) • For vocabularies and ontologies (RDFS, OWL) • For connect and querying (SPARQL)

II. How find OpenCourseWare? Searches based on Google Search Engine; Categories and tags.

In sum, the OCW searches have the following problems: * The query returns few relevant results compared with the number of retrieved irrelevant results. * Results are unsatisfactory because the search engine compares words and does not take into account the semantics of a term. * Results are simple, that is, do not combine retrieved results that are stored at different Websites. * The user has to extract the data and information from the located pages containing relevant results, because information is not available to machine agents in a machine-processable format.

6.613 OCW 65 institutions 12 languages [Dic 2011] Search Courses:

Advanced Course Search Browse by Language Browse by Source OpenCourseWare Websites Course Catalog (BETA) http://www.ocwconsortium.org/

1.126 associated universities, 23 Iberoamerican countries. 14M of teachers and students 1582 OCW couses 41 OCW providers 5 languages Accces by: Knowledge areas Authores Keywords Universities http://ocw.universia.net/

II The Value of the Semantic Web in Open Academic Initiatives Education empowered by Semantic Web Semantic Web technologies can also help to integrate the work of disperse institutions producing diverse data. The Linked Data aid the discovery, reliable re-use of data, provide improved provenance and facilitate automated processing by increased flexibility to changes in presentation and reduced ambiguity.

Challenges on management of OCW information generated and shared by Organizations (1) Large amounts of unstructured, and semistructured data. (2) Although the collected data from OCW repositories may have certain structure accepted by community, but not all data have an similar or compatible structure and meaning. (3) Open education materiales are shared as Information Silos or "Walled Gardens"

III. USING LINKED DATA ON OCW from Web of Documents

from human to human

to Web of Data

Discovery, Access, and Usages of Resources in the Web

General Framework for Publishing Open Educational Contents as Linked Data

Cycle to OCW to RDF Publication Monitoring for new OCW organizations and courses

Enrichment Linked OCW Data Repository

LOCWD Triplestore

Agent to include new OCW Organizations

RDF data

A new OCW organization

OCW Directories Listener

URI links A new OCW

Linked OpenCourseWare DataSet

Agent to include new OCW from universities stream of html content

Map the terms mined to terms already in the LOD Cloud

Connect OCW Data with Other RDF Repositories

URIs for OCW things RDF for describe OCW resources Links to other LOD - things

OCW Repositories Listener

LOCWD Linked Open Course Ware Data

LOERD Linked OER Data

LUD Linked Universities Data

RDF vocabularies

Extraction of OCW data Legend

Terms mined as RDF tripletes

RDF triplestore

Content extraction from HTML pages

Extraction of content from each OCW page

Extraction of data patterns (Classification, and stream of applying of clustering extracted content SNA techniques )

raw content

data corrected

Cleasing Data (detecting and correcting corrupt or inaccurate data

Relational database Software agent Get information from RSS subscription Get RDF content, if available

(CC license verified ) Use of crawling and scraping techniques

non-reliable data or erroneous data

Temporary Repository for store of html content extracted

Apply scraping technique Get embebed content in HTML pages

11

2. Common Vocabulary Modeling university name

University

university oficial web site OCW repository name

OCW repository

state of repository Platform

country

URL OCW repository RSS link

OCW001

Course title

Knowledge Area

Course Description Creation date Language

Tag list Tag Meaning Language

FirstName

OERs

tag

Licensed

Author

LastName Gender

OER link OER Subject OER type

university organization unit DBLP

OER language

12

Data available from an OpenCourseWare OCW University knowledge area

Title Author(s) Department syllabus bibliography year ects credits time autoself

description

Consuming and visualization of OCW-RDF Demonstration of Linked Data Queries in LOCWD: Queries, Maps, Mobiles, Recommender Systems, Faceted Searchs. Query A: Title for the OCW UPMSW08 PREFIX dc: PREFIX xsd: PREFIX locwd: SELECT ?ocwTitle, ocwDescription WHERE { dc:title ?ocwTitle. dc:description ?ocwDescription. }

Results >> Ontologies and Semantic Web, 2008 >> "The general objective is to provide students with a sound grounding of scientific, methodological and technological fundamentals in Ontological Engineering and the Semantic Web areas....

15

GoogleMaps to visualize Linked OCW Data

16

concept extraction desambiguation

entity equivalence You might like...

LUD publication

RDF Data Store

recomendations

Other OER OCW suggested

Recommender System based on Linked OCW Data

STUDY CASE: FACETED QUERY OF OCW BASED ON LINKED OPENCOURSEWARE DATA

OCW and OER

raw data now! Linked Data is Data Interoperability The need for communication and interoperation between autonomous and distributed information systems is increasing with the increasing usage of the Web.

e.g. interoperability between heterogeneous and distributed OCW/OER repositories

Benefits Why publish Linked OCW Data? • Because LinkedData holds the potential to move our OCW collections out of their silos • Open the data and content silos, to leverage the knowledge capital represented by our OCW repositories • To enrich our information landscape, to improve visibility • To improve ease of discovery open academic resources • To improve ease of consumption and reuse of OCW • To reduce redundancy in searched of OCW • Promoting innovation and Added Value to Open

Thank you for your Attention! 2012 World Open Educational Resources Congress Wednesday 20 – Friday 22 June, 2012 Room XII, UNESCO HQ, Paris, France

@nopiedra #WorldOER #OpenEducationalResources #OpenCourseWare #linkeddata #ocw #oer #SemWeb #SemanticWeb #LOCWD #LOD this work is licensed under a Creative Commons Atribución-NoComercial-SinDerivadas 3.0 Ecuador License http://creativecommons.org/licenses/by-nc-nd/3.0/ec/