Techniques, Technologies and Patents Related to Intelligent Educational Systems - E-Infrastructures and Technologies for Lifelong Learning - page 8

necessary (Murray, 1999). Tools/methods
recommending necessary changes are also
required. Data from learner-IES interaction
could be exploited for this purpose (Castro
et al., 2007).
• Support for domain knowledge construc-
tion/maintenance by multiple collaborat-
ing authors. Web-based technologies en-
able creation of author communities and
therefore multiple authors may collaborate
for construction/maintenance of domain
knowledge (Nesic, Gasevic & Jazayeri,
2008). Tools/methods facilitating collab-
orative domain knowledge construction/
maintenance are required.

the scope of the learning subject such as
experience in relevant fields, experience in
using computers, familiarity with the sys-
tem, etc.
• Knowledge level : Learner's knowledge
about the learning subject constitutes the
most important learner characteristic.
Estimation of knowledge level is an impor-
tant problem in user modeling.
• Learning style : It involves learner's pre-
ferred way of learning. Most learners fa-
vor particular approaches of acquiring and
processing information. An IES should be
able to assess learning styles of learners
and tailor instruction accordingly.

User Modeling Issues Representation of learner knowledge is closely

related with learner evaluation, i.e. a way to
estimate how well a learner has learnt specific
knowledge concepts. Although learner evalua-
tion is often used as part of the user modeling
unit, we consider more natural to include it in the
pedagogical module. There are a number of ways
for representing knowledge level in a user model.
The overlay model (Polson & Richardson,
1988) is a popular and simple way of representing
learner knowledge based on the pedagogical struc-
ture of knowledge domain. Learner knowledge is
considered to be a subset of expert knowledge. For
each domain knowledge item, the model retains a
Boolean/scalar value representing learner knowl-
edge level. For this purpose, fuzzy values may be
used. Learning process carries on until learner
knowledge is identical to expert knowledge.
A disadvantage of overlay model is inability to
represent possible learner misconceptions. For this
purpose, the buggy model (Polson & Richardson,
1988) is used which represents learner knowledge
as the union of a subset of domain knowledge and
a set of misconceptions. The buggy model provides
great assistance in correcting learner mistakes.
A general way of modeling learners is to
use stereotypes . Stereotypes denote predefined
classes/categories/groups of learners. Stereotypes

The user (or student) model records learner data
required to provide adaptation. The user model
should be neither incomplete nor too complex,
and so choosing the most appropriate learner
characteristics to record is a challenge. Primary
learner characteristics recorded in user model
(Brusilovsky & Millan, 2007; Polson & Richard-
son, 1988) are the following:

• History of learner's interaction with the
system : It may contain information such as
domain knowledge parts accessed by the
learner, answers given to questions/exer-
cises, requests for hints/assistance, brows-
ing activity, time consumed in studying
presented learning content, etc.
• Goals : Goals can be discerned to high-lev-
el goals, e.g. a concept, or low-level goals,
e.g. problem solving goals.
• Preferences : They mainly refer to edu-
cational content presentation parameters
such as multimedia type preferences (e.g.
text, images, or animations) regarding pre-
sented learning units or their level of detail.
• Background and experience : Learner's
background refers to experiences beyond