Folksonomy-based Course Authoring for Flexible Student
Modeling
Sergey Sosnovsky,Michael Yudelson
PAWS@SIS@PITT
Outline
• Definition: Folksonomy• Project Motivation • KnowledgeTree:
– Interface:• SEDONA
– Model Storage and (de)Composition• CUMULATE
– Flexible Student Modeling
• :– “A folksonomy is an Internet-based information
retrieval methodology consisting of collaboratively generated, open-ended labels that categorize content such as Web pages, online photographs, and Web links”.
» Thomas Vander Wal
• The main difference from formal knowledge models like ontology:– Subjective view (community-based)– Uncontrolled vocabulary– Poor structure
Folksonomy
Motivation:QuizGuide – successful example of topic-based navigation
• QuizGuide – provides adaptive navigation support based on fuzzy knowledge units – topics.
• Topic - …
Motivation:Customizable topic-based course structures
• Topics provide:– Good enough source for adaptation – Natural way to organize learning material– Easy way to index learning material
• However a topic-based model of the course as all folksonomic structures is subjective
• Main idea: To provide a friendly authoring interface for teachers for customization of their own topic-based course structures
• To scaffold topic reuse and topic-based modeling of student knowledge as well to enable inter-folksonomy knowledge transition we store topic models on the ontology server as rdf-documents
SEDONA: Model Storage and (de)Composition
• Stores three kinds of rdf-models:– Domain model:
• Topics and relations between them (if any)– LO repository (QuizPACK, WebEx) models:
• Metadata of LO’s– Basic Course structure
• Course metadata• Topic-LO relations
• On-demand:– retrieves models, combines them and reports the enhanced
model to the KnowledgeTree– receives enhanced model, decomposes, stores changes (in the
domain model and/or course structure)– reports to CUMULATE topic-LO associations
CUMULATE: Flexible Student Modeling
• Retrieving topics and their associations with LOs from SEDONA and cashing it
• Student knowledge inference in new topics based on student activity with LO
• Reporting of student knowledge to adaptive services
Conclusion
• Framework for:– Easy authoring of topic-based course
structures– Automatic modeling of student knowledge in
newly created topics• Using of ontology server scaffolds topic
reuse and allows inter-model knowledge inference.
Topics• Provide useful way of learning material
aggregation• Play two roles for adaptation:
– As domain elements used for recourse indexing and students’ knowledge assessment
– As interface elements used for resource structuring and navigation
• Coarse-grained (not precise UM)• Subjective
http://www.sis.pitt.edu/~paws/