(Personalization of learning material in web-based education)

1
(Personalization of learning material in
web-based education)
Personalisering av læringsinnhold i e-læringskurs

• Håvard Narvesen 05HMTMT

2
Overview

• Employer Apropos Internett (Hamar, Norway)
• Main task Study ways to adapt learning material
  based on individual competence gaps
• Supervisor Rune Hjelsvold
• Keywords E-learning, personalization, adaptive
  hypermedia

3
Introduction
What is a Learning Management System (LMS)? What
is the problem with presentation of most
web-based education material today? How can
personalization improve web-based education?
4
Problem area
One-size-fits-all-scenario
Personalized material
5
Why personalize learning material?

• It makes web-based courses more relevant to each
  learner.
• By making e-learning courses adaptable to each
  learners pre-knowledge, learners may start the
  same course at different entry levels.
• If the learning material doesnt feel relevant,
  then the learners motivation weakens. Audun
  Gjevre, Apropos Internett

6
Research questions

• S1 Hvilke egenskaper bør et nettbasert
  læringssystem inneha for å støtte personalisering
  av læringsinnhold basert på hver kursdeltakers
  kompetansegap?
• S2 Hvilke er de største tekniske utfordringene
  ved implementasjon av et adaptivt e-læringskurs,
  der innhold tilpasses basert på kursdeltakerens
  forhåndskunnskaper?
• S3 Hvordan oppfatter kursdeltakerne
  automatisert pretesting?

7
Method

• S1 A literature study and an interview with an
  expert was used to understand relevant concepts
  and point out key characteristics of educational
  adaptive learning systems.

• S2 A prototype of a system, capable of
  personalizing learning material, was build in
  order to bring out major technical difficulties.

• S3 An experiment was carried out to get feedback
  from a set of learners on implemented
  personalization techniques. Qualitative and
  quantitative methods were used to gather data.

8
Some results Study of characteristics (S1)

• By pre-testing each users knowledge prior to the
  web-based course, it is possible to unveil human
  competence gaps, and let them influence the
  personalization.
• The pre-test cannot be too resource-demanding
  neither for teachers or learners.
• Computer agents are commonly used to support
  learners in modern web-based educational systems.

9
Some results Technical challenges (S2)

• Describing and dividing learning material suited
  for personalization. The SCORM standard is not
  perfectly suited for advanced personalization.
  (Abdullah et al., 2003)
• Building automated pre-tests, and then evaluate
  the results
• Automatically adapt learning material to each
  learner based on results from the pre-test and
  the learning goals. (Knowledge based)
• Implementation of agents for supporting
  adaptation ? one learner many teachers

10
The experiment

• A test group of 11 learners used the prototype to
  carry out a web-based course.
• The course concerned computer viruses.

• A simple pre-test determined the available
  learning material.

11

• The structure of the course

• The pre-test was organized as follows

• This means that the pre-test consists of the
  users pre-knowledge for each of the main topics
  in the course. The pre-knowledge was included as
  a part of a user model.

12
Some results Experiment (S3)

• All participants agreed to spend 5 or more of
  the total time a course demands in order to
  personalize a course (in the future).
• Only 2 of the 11 learners fully agreed with the
  technique for filtering learning material
  implemented in the prototype. These results
  confirms conclusions from other researchers that
  creating a system that can predict every learners
  competence gap with 100 accuracy, is
  unrealistic.
• Also, the learners view on Personalization in
  e-learning, how they like to be tested, how they
  liked link-personalization and more.

13
General conclusion (preliminary)

• The experiment in this work, and other studies,
  suggest that a pre-test should be used to decide
  which learners that need (or not need) extra
  attention, rather than entirely delimit the
  course material.
• Most test-learners did not like that the system
  totally decided what they should read and not.
  Based on information from the learners, the
  pre-test results should rather be used to make a
  suggestion of what to prioritize in the
  e-learning course.

14

• Thank you for your attention!
• Any comments or questions?