Positioning in Learning Networks

PROJECT: Positioning in Learning Networks

Website: http://www.onderzoekinformatie.nl/en/oi/nod/onderzoek/OND1308428/
(temporally)

Project description:
Distributed, self-organized learning networks for lifelong learning to which multiple providers contribute
resources require new techniques to determine where learners can be positioned in these networks.
Positioning requires us to map characteristics of the learner onto characteristics of learning materials and
curricula. Considering the nature of the network envisaged, maintaining data on these characteristics and
ensuring their integrity are important and complex tasks. In this project, positioning is based on an
abstract semantic domain model generated from the data contained in the learning network with the use of Latent Semantic Analysis (LSA). LSA allows to generate a common semantic framework for characteristics of the learner, learning materials and curricula.

Before a learner can navigate along a path, we have to decide upon a point from where the
learner might start. If, for example, a learner has already reached the level of competency that is addressed
in some of the nodes he/she may be advised to skip these nodes. Positioning in Learning Networks addresses the
question what the learner’s position within the network is, considering the learner’s needs and goals as well
as his/her competency levels. We refer to this as the positioning question. The positioning process
establishes an entry point in the network with respect to one learner goal. The process uses data on the
learner’s competencies to establish which steps of a plan to reach a goal need to be followed by the learner
and which steps can be exempted.

Objectives:
Primary objectives
• Demonstrate and test the quality and technical feasibility and scalability of positioning using text-based
similarity measures using LSA .
• Develop and test a prototype positioner in the LN software context.
Secondary objectives
• Evaluate the reliability and validity of positioning using LSA by comparing computer-generated decisions
to human expert decisions.
• Compare and contrast applications of current practice in exemptions and Accreditation of Prior Learning
to computer-based positioning.
• Develop guidelines on the use of LSA parameters for positioning.
• Specify requirements for prototype positioner.
• Develop positioning prototype in an LN context.
• Test prototype positioner.

Outcomes:
• A validated prototypical positioning system that is integrated and tested in the Learning Network
Infrastructure that will be developed in the project 'Learning Networks Integrated'.
• A validated method (guidelines) for the positioning of learners in learning networks on the basis of text
similarities. These guidelines will identify requirements for the materials to be used as input to define a
domain model as well as the requirements for the learning materials used in the similarity calculation.
The guidelines will specify the parameters to be used for the underlying computations (parameters
cover a wide area, ranging from the type of frequency measure to vector normalization). These
guidelines will be developed tested and refined through a series of iterations in all phases of the
project.
• A validated specification of the use of LSA-based positioning systems within learning networks,
including indicators for reliability and validity of LSA based positioning.

Project leader and researchers involved:
Project leader: Jan van Bruggen
Team members: Ellen Rusman, Bas Giesbers

Project partners:

Planning (including current status):
Begin date:
End date:

Current status:

Relations with other projects within the Development Programme:
ASA: Cooper will build on work on community formation started in the ASA project
ROMA