Collaborative Filtering

1
Collaborative Filtering Initiatives in the
Social Sciences
Presentation The Great Plains Network
Conference Building the Collaborative
Infrastructure Bridging the Science-Technology-F
unding Divide November 15-16, 2000 Linda E.
ODonnell Center for Research on
Learning University of Kansas
2
Learning Variables Research

• Learning as Change
• 11,000 variables
• 33 domains
• 8 categories
• Change at point in time
• Change points across time

3
Human Learning

• Learning vs. Teaching
• 1 - Year Review of Research in 7 Areas
• 20 Years of Clinical Differential Diagnosis
• Research Synthesis of 25 Yrs.
• Reading, Writing, Spelling, Mathematics
• Language Communications
• Social, Emotional, and Affective Behavior
• Cognition Perception/Learning Intelligence

4
Interdisciplinary CollaborationJuries
Coordinators of 4 Juries of National
International Teams of Experts in the Field

• Biological systematics and ecology
• Steven Ashe, Biological Science, KU
• Mathematical modeling and graphic imaging
• Estele Gavasto, Mathematics, KU
• Information technology with bioinformatics,
  database structures, and Internet 2
• Carl Kurt, Engineering, KU
• Learning variables research
• Daryl Mellard, CRL, KU

5
How collaborative filtering works
Other Individual Preferences
Synthesis of Preferences
Recommendation based on collaborative
filtering (Martin, 1999)
Individual Preference
6
Mapping the Learning Variables

• Taxonomies
• Biological variation versus disabilities
• Scientific versus standards based failures
• Features analysis
• Onset, duration, frequency, co-morbidity,
  intensity, BvC, critical feature identification
• Differences analysis
• Data mining with bioinfomatics
• (e.g., reverse Markov method)

7
Study of Learning the Elusive A x TI

• Disciplines of Inquiry for Research in Education
• Arts-based methods
• Historical methods
• Philosophical inquiry
• Ethnographic methods
• Case study methods
• Survey methods
• Comparative experimental
• Quasi-experimental
• (American Educational Research Association,
  Jaeger, 1997)

Clinical Studies
Innovative Programs
Product Development
Basic Research Precommercial RD
8
Students Response to Levels of Increasing
Difficulty
Conditions Of Prediction
On Students Learning Curve
On Tests
On Lessons
On Tasks
- - - - -
- - - - -

Too Hard no learning increment

All items or tasks unknown
Frustration level
III.
Ceiling
- - - -
- - - -
Just Right optimal learning related to of known
unknown information
Some known and some unknown


Basal Versus Ceiling (BC) Range
Instructional level
II.




Independent level
All items or tasks known
Too Easy no learning increment
I.
Basal
9
Innovations Diffusion

• Crossing the Chasm
• Innovators
• Early adopters
• Mainstream users
• Late adopters

(Moore, 2000, p. 141)
10
Technology Adoption Life Cycle
(Moore, 2000, p. 143)
11
Internet 2

• Collaboration
• Collaborative filtering via Internet 2
• Data Analysis
• Real time data gathering (cc. Weather)
• Test bed data bases (cc. Murray Center)
• On-site vs. remote data gathering (retired
  cohort)
• Cycle from classroom to university researchers

12
4 Technology Transfer Effectiveness Models

• Out-the-door model of Technology Transfer
• Market Impact Model of Technology
• Market Definition of Effectiveness
• Political Definition of Effectiveness
• TTE is only one of many technical activities
  and
• maybe not the most important (Bozeman Crow,
  1998)

13
Innovations Evolution

• Group interaction
• Communication of 2 or more
• Any meeting
• Group work leading to individual output
• Group work leading to group product
• Collaborative filtering

14
Scientific Study of Learning

• Using scientific research methods for study of
  social sciences data
• Using social sciences research methods for the
  study of scientific data (DLI for engineering,
  physics, computer science)

15
Consilience for Social Sciences?

• General (no)
• Specific to Learning (yes)
• Unity of knowledge (Edward O. Wilson)
• Useless effort (Feynman comparing natural
  sciences with social sciences)
• Principa as natural philosophy (Newton)

16
Hard, Harder, Hardest Sciences

• Natural sciences
• Physics
• Engineering
• Astronomy
• Biology
• Learning
• Others

• Social sciences
• Sociology
• Medicine
• Economics
• Anthropology
• Education
• Others

• Mathematic
• Scientific method
• Technology
• Others

17

• Hard or natural sciences
• Harder sciences (the concilience challenge)
• Hardest science
• Press of human services
• Crisis delivery for social and educational areas
• Multiple variables simultaneously co-mingling

18
Coming of Age with Internet 2

• Science of Learning, not just science, math,
  and technology learning
• Maturation of knowing silicone snake oil and
  chasm of innovation diffusion versus focused
  appropriate areas
• Coming of age of the scientific, technological,
  and collaborative methods

19
The Canary or the Dovenot the only choice

• Sailors or miners canary
• Sailors dove
• lest you say it is for the birdsconsider
  http//kanCRN.org -25 states 3 countries in the
  analogy

20
The KU-CRLs Six-Stage Research ProcessDonald D.
Deshler, Director of CRL
21
Divisions within the Center for Research on
Learning (CRL), University of Kansas, Lawrence,
KS (Revised 5/22/00).
22
Peer out into the new science of Learning
Variables Research
LAtmosphere Meterologie Populair, 1888,
Deutsches Museum, Munich, Germany
23
Collaborative Filtering
Other Individual Preferences
Synthesis of Preferences
Individual Preference
Recommendations based on collaborative
filtering (Martin, 1999)