PewePro 2

Automated Domian Model Creation

A method for automatization of adaptive course metadata creation.

Overview
Inputs: Learning objects content, learning object links, social annotations
Outputs: Adaptive course domain model

Addressed Problems

We focus on domain model authoring support, which is crutial for teachers as content authors. The amount of semantic descriptions necessary for an adaptive course is so high that it is almost impossible for a human to create it and maintain it manually. The situation is even worse in the case of user generated content during learning (e.g. comments, tags) that has to be assigned with metadata with no human help.

Description

The method consists of the three major steps (see Figure 1):

  1. resources preprocessing
  2. relevant domain term (RDT) extraction, and
  3. relationship discovery

In the first step, resources are preprocessed in order to compose representation for further processing. The resources (both learning objects and social annotations created by learners) are analyzed and extended vector representation is created. In the second step we select the most relevant terms in each learning object representation and create resource-term associations. Relationship discovery step is very important step for overall domain model "feasibility" in adaptation process. We perform graph, linguistic and statistical analysis to discover two types of relationships: relatedness and is-a relationships.

Automatic domain model creation

The method for automatic domain model creation.

There also is the fourth (optional) step - domain model finalization, where a teacher eventually modifies a created course according to his needs; making the whole approach automated

Publications

  1. Ċ imko, M.: Automated Domain Model Creation for Adaptive Social Educational Environments. In: Information Sciences and Technologies Bulletin of the ACM Slovakia. ACM Slovakia Chapter. Vol. 3, No. 2., 2011, pp. 119-121.