An Adaptive ELearning Service for Accessing Interactive Examples

1
An Adaptive E-Learning Service for Accessing
Interactive Examples

• Peter Brusilovsky, Michael Yudelson, Sergey
  Sosnovsky
• TALER Lab, University of Pittsburgh, USA

2
Overview

• Background
• WebEx
• NavEx
• Demo
• Indexing code examples for NavEx
• Annotating examples
• Architecture
• Conclusions

3
Background

• Problem
• Undergraduate programming course
• 2-5 C-code examples presented at each lecture
• Students cannot grasp examples during on the
  lecture
• Off-lecture the examples posted online lack
  teachers comments
• Attempted solution
• Provide online examples that are supplied with
  teachers comments

4
WebEx

• WebEx Web Examples
• Individual lines of code are supplied with
  teacher comments
• Launched in 2001

5
WebEx contd
6
WebEx usage data

• Users 135
• Mean number of click per user 75.6 (Max - 1708)
• Mean length of session 13.9 (Max - 347)

7
WebEx usage data contd

• Mean number of examples per session - 2.4 (Max -
  34)
• Mean number of examples per user - 9 (Max 34)

8
WebEx problem

• Currently there are has 60 examples (3-5 per
  lecture)
• Students tend put off reviewing lecture examples
• After 1 lecture 4 examples, 2 lectures 8,3
  lectures 12
• At certain point theres a lot to review
• Problem lack of guidance as to which examples to
  review and at what time
• Attempted solution
• Introduce adaptive annotation of examples to
  provide guidance

9
NavEx Interface
10
Demo

• NavEx

11
NavEx interface revisited
12
NavEx interface revisited
13
NavEx interface revisited
14
NavEx interface revisited
15
NavEx interface revisited
16
NavEx interface revisited
17
NavEx Adaptive navigation support

• NavEx uses traditional navigation support based
  on prerequisite relations (AHA, InterBook,
  ELMArt)
• This kind of navigation support is known to be
  effective
• It requires content to be indexed
• Indexing by hand is (a) time-consuming and(b)
  prone to error

18
NavEx Adaptive navigation support

• Problem
• Allow new examples to be added w/o manual
  indexing
• Attempted solution automatic indexing of the
  examples
• Indexing examples with concepts
• Ordering examples by dividing concepts into
  prerequisites/outcomes

19
Structuring code examples for NavEx

• Indexing examples with concepts
• Concepts derived from C language constructs
• C-code parser (based on UNIX lex yacc)
• 51 concepts used (in 60 examples)
• e.g. include, void, main_func, decl_var, etc
• At this stage no distinction between concepts
  made (as prerequisite/outcome )

20
Structuring code examples for NavEx

• Divide concepts of each example into
  prerequisites/outcomes
• Examples are grouped into lectures (by teacher)
• For each example
• Concepts, corresponding to examples from previous
  lectures are considered prerequisites
• All other concepts for this example are outcomes
• Examples are connected through concepts

21
Structuring code examples for NavEx

• Concept indexing Concept division

22
Annotating examples

• Example annotation is based on overlay user model
• User progress with ith example ?0,1
• Progress changes as user explores more
  annotations to lines of C-code (clicks on green
  bullets)

23
Annotating examples

• Annotation
• Each example is supplied with an icon
• example is not ready to be explored
• example is ready to be
  explored, plus user progress is displayed
• example availability is computed based on
  prerequisite concepts and user progress
• filling reveals amount of users work with example

24
NavEx Architecture
25
Conclusions

• NavEx shell for accessing annotated C examples
• NavEx offers prerequisite based guidance
• what examples are ready to be explored
• what is the students current progress with the
  example
• NavEx is currently being tested
• Evaluation is yet to be done

26
Contact information

• Michael V. Yudelson
• School of Information SciencesUniversity of
  Pittsburgh
• E-mail mvy3_at_pitt.edu
• Web http//www.pitt.edu/mvy3