ReMashed – Recommendation Approaches for Mash-Up Personal Learning Environments in Formal and Informal Learning Settings at MUPPLE09 in ECTEL 2009, Nice, fr
Ähnlich wie ReMashed – Recommendation Approaches for Mash-Up Personal Learning Environments in Formal and Informal Learning Settings at MUPPLE09 in ECTEL 2009, Nice, fr
Ähnlich wie ReMashed – Recommendation Approaches for Mash-Up Personal Learning Environments in Formal and Informal Learning Settings at MUPPLE09 in ECTEL 2009, Nice, fr (20)
ReMashed – Recommendation Approaches for Mash-Up Personal Learning Environments in Formal and Informal Learning Settings at MUPPLE09 in ECTEL 2009, Nice, fr
1. ReMashed – Recommendation Approaches
for Mash-Up Personal Learning
Environments in Formal and Informal
Learning Settings
Hendrik Drachsler, Dries Pecceu, Tanja Arts, Edwin Hutten, Peter
van Rosmalen, Hans Hummel & Rob Koper
2. Personal Environments
Nowadays …
More Information
Blog Reader Providers
Social
Bookmarking
Various
Communities
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 2 | September 28, 2009
4. Related Work
• Pedagogical Scenarios
Formal Learning (e.g. Hermans & Verjans, 2009)
Informal Learning (e.g. Wild, Kalz & Palmer, 2008)
• Use Case Studies
4 experiments (e.g Van Harmelen, 2008)
• Technology Development
Language design for a PLE
(e.g. Moederitscher, Wild, Sigurdarson, 2008)
Widget Interoperability
(e.g. Sire, Vagner, 2008)
http://mindmeister.com/15237440/r-d-on-mupples
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 4 | September 28, 2009
5. Selection problem because …
…of the amount of
data that is
emerging in
MUPPLEs.
…learners can be
overwhelmed by
the plethora of
information.
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 5 | September 28, 2009
6. Can we create a Recommender System
Today, Recommender Systems
for MUPPLEs? our decisions
supporting
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 6 | September 28, 2009
7. What is ReMashed?
A Mash-up environment that allows you to
personalize emerging information of online
communities with a recommender system.
You tell what kind of Web 2.0 services you
use and then you are able to define which
contributions of other members you like
and do not like.
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 7 | September 28, 2009
8. Goals for ReMashed
1. End-User level
Providing a recommender system for Web 2.0
sources of learners in MUPPLEs.
2. Researcher level
1. Offering researchers a system for the
evaluation of recommendation algorithms for
learners in MUPPLEs.
2. Creating user-generated-content data sets for
recommender systems in MUPPLEs.
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 8 | September 28, 2009
11. The 2nd Release
DUINE Prediction
Engine
Database User Interface
of Items
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 11 | September 28, 2009
12. How does it work?
ReMashed uses collaborative filtering to
generate recommendations.
It works by matching together users with
similar tastes (neighbours) on different
Web 2.0 resources (delicious, Flickr, blog
feeds, Slideshare, Twitter, and YouTube).
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 12 | September 28, 2009
13. How does it work?
Cold-Start = Tag-based recommendation
Collaborative Filtering with ratings
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 13 | September 28, 2009
15. Context Variables
Formal Learning (Top-down environments)
Curriculum (Closed-Corpus)
Teacher directed
Predefined learning resources, learning goals
Maintenance
Informal Learning (Bottom-up environments)
More self-directed learning goals
Responsible for own learning pace / path
Learning resources from different providers (Open-Corpus)
Lack of maintenance
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 15 | September 28, 2009
16. Recommendation Approaches
Learning settings, environmental conditions
and the task greatly affect the design of
recommender systems in TEL.
systems in TEL.
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 16 | September 28, 2009
21. You can use it as well!
Please sign up at:
Register at ReMashed
remashed.ou.nl. starts mashing.
http://remashed.ou.nl
Enter your favorite Taste your
Web 2.0 potatoes. personal
flavor of
Web 2.0.
Join the
community.
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 21 | September 28, 2009
22. Many thanks for your interest!
This slide is available here:
http://www.slideshare.com/Drachsler
Email: hendrik.drachsler@ou.nl
Skype: celstec-hendrik.drachsler
Blogging at: http://elgg.ou.nl/hdr/weblog
Twittering at: http://twitter.com/HDrachsler
hendrik.drachsler@ou.nl
EC-TEL 2009, 2nd MUPPLE workshop, Nice, FR
Page 22 | September 28, 2009
Hinweis der Redaktion
A solution towards this problem might be recommender system technology. The main purpose of recommender systems on the Internet is to pre-select information a user might be interested in. For instance, the well-known company amazon.com [8] is using a recommender system to direct the attention of their users to other products in their collection. The motivation for a recommender system for Mash-Up Personal Learning Environments is to improve the ‘educational provision’; to offer a better goal attainment and to spend less time to find suitable learning material. Therefore, we developed a recommender system that offers advice to learners based on their Web 2.0 resources regarding the most suitable learning materials to meet their individual competence development.
Organization can pre-structure and control available learning goals, knowledge domains etc. Maintenance effort, design activities needed before the runtime , Ontologies Metadata Predefined learning paths
It builds up a hierarchy of items by continuously merging the two most similare items / groups into a new group Measure of most frequently used keywords for each blog posting by simp Open Corpus Absence of maintenance and structure Metadata Predefined learning paths Nearly no maintenace improve through emerging behaviour le using words counts or Reuters service