PewePro



Student Model Updater

A method for creation and maintenance of learner characteristics on the basis of their behavior in the course.

Overview
Technologies used: Java, XML (JAXB), Sesame repository
Inputs: User model, events reported from presentation layer – learning object reading time, feedback from currently displayed learning object (knowledge and interest)
Outputs: Changes of characteristics in user model

Addressed Problems

Learners have different knowledge, interests and goals and thus require personalized learning content. Personalization consists of two processes – estimation of user’s characteristics in a user model and adjust (or adaptation) of the educational content according to modeled user’s characteristics. We propose a novel method for maintenance of user’s characteristics based on spreading a change.

Description

Our domain model consists of two parts: concept space and learning object space. We employ the standard overlay user modeling technique, which allows us to express user characteristics related to both concepts and learning objects. A user characteristic related to a particular concept determines user characteristics for all learning objects which falls within the scope of the concept. That means, that we are able to estimate and maintain learner’s characteristics also for the learning objects which have not been visited by the learner yet.

While a user works with a learning object, we obtain her characteristics for that learning object (e.g. estimated interest). Since there is a connection between learning objects and other parts of the domain model, we spread a change to other, related parts of the domain model. Our user modeling approach consists of following steps (see figure below):

  1. Setting characteristics for current learning object (e.g. Example with I/O operations) – by considering user’s activity within the learning object, we can find out user’s interest and knowledge for that learning object.
  2. Spreading changes of characteristic’s values from current learning object to related concepts (concepts Display and Scan).
  3. Spreading changes of characteristic’s values in concept space (from concetps Scan and Display to other related concepts: Input/Output and Files).
  4. Spreading changes of characteristic’s values from concepts where some characteristics have been changed to related learning objects (Example with writing to file and Example with displaying items of array).

structure of proposed domain model.

Sequence of steps in process of change of characteristic