16
Discovering occurrences of user- defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František Babič Ján Paralič KKUI, FEI TU Kosice

Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Embed Size (px)

Citation preview

Page 1: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Discovering occurrences of user-defined patterns in historical data representing

collaborative activities in virtual user environment

Jozef Wagner

František Babič

Ján Paralič

KKUI, FEI

TU Kosice

Page 2: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Outline

Background Source data Analysis Patterns Conclusion

WIKT 2010, Bratislava, Slovakia 2/16

Page 3: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Background (1)

Virtual user environment as working or learning platform

Collaborative activities within various groups of users.– Lead to new learning and knowledge practices

All activities within virtual user environment are logged

WIKT 2010, Bratislava, Slovakia 3/16

Page 4: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Background (2)

Complex processes, Ill defined and not easily formulated

Not all activities are captured– F2F meetings, field trials, minutes, interviews

Researchers and Teachers wants to evaluate user behavior during these activities, identify and understand knowledge practices

WIKT 2010, Bratislava, Slovakia 4/16

Page 5: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Source data

Dedicated server accepts event logs from various end user tools participating in the learning process

In each event– Identify user, object, working group– Type of action, Time– Custom properties

WIKT 2010, Bratislava, Slovakia 5/16

Page 6: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Analyses

Quantitative– Summarized information– Aggregation, filters

Social network analyses Historical retrospective

– Based on timeline views

WIKT 2010, Bratislava, Slovakia 6/16

Page 7: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Quantitative analyses

WIKT 2010, Bratislava, Slovakia 7/16

Page 8: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Timeline view (1)

WIKT 2010, Bratislava, Slovakia 8/16

Page 9: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Timeline view (2)

WIKT 2010, Bratislava, Slovakia 9/16

Page 10: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Timeline view (3)

Visualization of performed events on the timeline in chronological order

Interactive tool Filtering, zooming Possiblity to add external events and

annotations Posibility to create and use patterns

WIKT 2010, Bratislava, Slovakia 10/16

Page 11: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Patterns (1)

Process discovery techniques are not applicable Our processes are not rigid, well defined Patterns help to identify critical points and best

practices Attempt to formally and explicitly define parts of

the process Interactive and iterative

WIKT 2010, Bratislava, Slovakia 11/16

Page 12: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Patterns (2)

Patterns is a sequence of pattern elements, each representing one generalized event

Include logical operations, unification Beyond simple sequence, branching Define weight, multiplicity Searching generates tree of matches

WIKT 2010, Bratislava, Slovakia 12/16

Page 13: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Clojure(.org)

LISP on JVM Better Java than Java Dynamic functional homoiconic language Lazy evaluation for sequences Concurrent programming made easy

– STM– Immutable data

WIKT 2010, Bratislava, Slovakia 13/16

Page 14: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Open Questions

Formal description for patterns Filter out similar results

– But provide means to display them, if requested

Return also partial matches

WIKT 2010, Bratislava, Slovakia 14/16

Page 15: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Database for logs

Mysql MonetDB (column oriented) Key/Value storage (CouchDB) Graph storage

– Neo4j (JSON)– Allegro graph (SPARQL)

WIKT 2010, Bratislava, Slovakia 15/16

Page 16: Discovering occurrences of user-defined patterns in historical data representing collaborative activities in virtual user environment Jozef Wagner František

Thank you for your attention

[email protected]

[email protected]

[email protected]

WIKT 2010, Bratislava, Slovakia 16/16