Social, Personalized Lifelong Learning - E-Infrastructures and Technologies for Lifelong Learning - page 111

Example: The type of a physical device at-
tribute can be “resolution”. The PDA resolution
is 240×160, and the computer screen resolution
is 1280×1024. Please note that when more users
are collaborating, the common denominator of
the different devices used by the different users
is the one that is selected, e.g., the minimum
resolution at which all collaborating partners can
view the item.
Definition 32: A user device link ul

generic environment model (as illustrated by the
physical device model, PDM , above).
Example: This type of definition is very ge-
neric, and allows for the actual implementation to
be done either by traditional rule-based systems, or
by actual mathematical formulas, or by Bayesian
networks, etc. An example is a group-based adap-
tation support via recommendations techniques,
such as recommended learning content (which
is rated high using Definition 4) based on the
learner's profile (which is represented in Social
User Model). This can be implemented with a
function based on the content (from the i(IM)) and
the rating (from the same model), and a personal
threshold for a given student for accepted rating,
stored in the i(UM). This will influence which
items will be shown: i(PM). This represents con-
tent adaptation as in scenario 1 (item or module),
scenario 3,5 (module) and Figures 4 (adaptive
reading) and 6 (adaptive authoring). Another
example is that of recommended expert learners
based on the learner's profile, as in the second part
of scenario 1, as well as in the recommendationof
a group member in scenarios 2 and 4. This maps
i(UM) to i(UM), and it actually means adding a
temporary link between the current user and the
recommended person in the user's user model.
In the context of MOT 2.0, the learner's
knowledge (experience level) is the main factor
in the recommendation process, as it reflects who
is expert in the selected domain. We recognize
two steps — the recommendation step and the
communication step. MOT 2.0 as presented in
this chapter focuses on recommendation, but not
on the communication between learners (which
is left to further research and implementation).
Further work on the communication step has
already started. Also various recommendations
based on cosine similarity between elements have
been added.

PDL
is a tuple <id uc , pm> where id uc is an identifier
for user uc, and pm is the presentation media.
Example: A user, say “Jonny”, can use different
devices, say “PDA”, “Desktop”, “Group device”,
etc. The latter is used in connection with the last
example in definition 33.

∈

Adaptation Model

Definition 33: The social adaptation model AM
is formed by the set of all adaptation maps ( AM
⊆ CM ), containing all information (resources and

links between them) of the adaptation (dynamic
changes) performed in a social adaptive system,
based on all other static models in the framework
( Social Resource Model , Social Domain Model ,
Social Goal and Constraints Model , Social User
Model , Environment Model - here: Physical Device
Model, Presentation Model) .
Example: The adaptation model is thus the only
dynamic model in the framework, and it uses the
other models as 'ingredients', to form the overall
'recipe' for social adaptivity. This model in itself
can have many components. An example is the
LAG model (Cristea & Verschoor, 2004), which
can be extended towards collaborative adaptation.
This is not further pursued in the current chapter,
due to lack of space.
Definition 34: An adaptation map AM

AM (or an adaptation strategy ) of the social
adaptive system (SAS) is a collection of mapping
functions f:{i( IM )*, i( DM )*, i( GM ), i( UM )*,
i( EM )*, i( PM )*} -> {i( PM )*, i( UM )*}, where
i(X) is an instance of X, and EM represents the

∈