PewePro 2

Tag Cloud-based Navigation

A method for creation adaptive keyword recommendation utilizing the user model, that supports the users's navigation on the site.

Overview
Inputs: User model, current learning object being read within the presentation layer
Outputs: Recommendation of a specified part of curriculum represented by a selected concept

Addressed Problems

Each student has different interests and knowledge so they browse different parts of curriculum. Our aim is to help the navigation on the site by displaying of personalized keyword (concept) recommendations which represent a specific topic. The focus is set on effective detection and elimination of shortcomings of students’ knowledge.

Description

Since the goal is to design the navigation using keywords, we chose a layout in a form of tag clouds. There are two categories of concepts displayed in the cloud divided into two boxes.

The first type is Related concepts - all the concepts linked to actually opened learning object. They serve for quick switching among similar learning objects belong here.

We recommend – the second part of the cloud. It is a personalized set of recommended concepts created by the help of user’s activity history.

The tag cloud has a specific orientation. It does not navigate the user to the concrete learning object but recommends topics he/she should deal with. If the user clicks on any concept in the tag cloud, the related learning objects will be highlighted in the main menu. In this way we indicated the proper direction for the user but the final decision to select one of the recommended objects is up to the user.

Recommendation methods are processes, which have certain inputs and outputs. Inputs of these processes are learning objects of a given type. The input object’s type depends on the character of the method. The output for one input learning object is one concept. Each output concept has an associated priority, which represents the strength of the recommendation. We select a set of concepts with the highest priorities, which are displayed in the tag cloud, part “We recommend”.

One of the recommendation methods is Method based on questions and examples solutions (see figure below). The input of this method is a question or example, which was incorrectly solved by the user. In this case the priority of the concepts linked to the input learning object is given by quotient of the representation relation weight and the knowledge of the concept. If there are more concepts linked to one input learning object, we choose the one with the highest priority. The recommendation outputs, are concepts of learning objects, which should be read over by the student in order to get the required topic under the control.

Another method is Method based on reading of learning materials (see below). The input is a text-explanation that was read (visited) by the user, but it doesn’t have any linked concepts, which have a knowledge relation with the user. It is about learning materials students have already read, but have not practiced yet. The system recommends questions and examples for the student, which are related to the read text.

Selecting concepts for recommendation

Selecting concepts for recommendation.

Publications

  1. Fejes, M: Concept-Cloud Navigation in Educational Web-Based System. In: Personalized Web - Science, Technologies and Engineering: 9th Spring 2011 PeWe Workshop Viničné, Galbov Mlyn, Slovakia April 1, 2011 Proceedings. Bratislava: STU, 2011, pp. 5-6.