• No results found

Current Adaptive E-Learning Systems 29

2. Research Background 16

2.5. Current Adaptive E-Learning Systems 29

This section gives an overview of some popular adaptive e-learning and authoring systems. The presented systems are InterBook (Brusilovsky et al., 1998), AHA! (De Bra and Calvi, 1998), KBS Hyperbook (Nejdl and Wolpers, 1999) and MOT 1.0 (Cristea and de Mooij, 2003a).

2.5.1. InterBook

InterBook (Brusilovsky et al., 1998) is one of the first adaptive systems for authoring and delivering adaptive electronic textbooks on the Web. It uses two types of knowledge, knowledge about the course being taught, and knowledge about learners. The domain model presents the structure of an adaptive electronic textbook. The domain model provides a structure for representation of the learner’s knowledge. For each domain model concept, an individual learner’s knowledge model stores some value which is an estimation of the student knowledge level of this concept.

All learners’ actions (page access, problem-solving, answering questions) are stored and used to update knowledge levels for the related concepts. Another component of the student model is the model of student's learning goals. In addition, each learner may have a learning goal. A learning goal is a set of concepts to be learned. According to Brusilovsky (Brusilovsky et al., 1998):

“A sequence of assigned learned goals forms an individual order of learning. Adaptive guidance mechanisms will ensure that the student achieves the first learning goal in a sequence, then the second one, and so forth.”

InterBook uses coloured bullets and changed font styles (different font size, bold fonts) to achieve adaptive navigation support via link annotation. In addition to links annotation InterBook provides direct help about the recommended next page the learner should access. Moreover, InterBook also uses navigation support.

2.5.2. KBS Hyperbook

The Knowledge Based Systems Hyperbook (Nejdl and Wolpers, 1999) is a tool for designing open adaptive hyperbook courses on the web. It was implemented in the Institut für Rechnergestützte Wissensverarbeitung, University of Hannover, Hannover, Germany. The open feature means the system is able to integrate distributed information resources. For example, the KBS Hyperbook was used for an introductory course “Introduction to Programming in Java” in the computer science department at University of Hannover, and

the course was integrated with pages from the Sun Java Tutorial into a hyperbook. According to Henze and Nejdl (Henze and Nejdl, 1999):

“the adaptation techniques used for this course are based on a goal-driven approach. This allows students to choose their own learning goals and to get suggestions for suitable projects and information units covering the knowledge required to reach these learning goals.

The learning content (domain model) of the hyperbook consists ofsemantic information unitsandproject units. Both of these refer to the real content to be displayed on the Web as pages of the hyperbook (Nejdl and Wolpers, 1999):

 SemanticInformationUnits: refers to information units about Java Objects. They are indexed by KnowledgeItems, which gives for each knowledge item one main information unit.

 ProjectUnits represent project descriptions, and are indexed by those knowledge items which the student needs to know in order to successfully work on these projects.

The learner model in the KBS Hyperbook system is based on a pedagogical component. This pedagogical component consists of the knowledge items and adds a partial sequence between these knowledge items, to describe learning dependencies. The KBS Hyperbook system uses the Bayesian network for modelling the learner's knowledge (Henze and Nejdl, 1998).

2.5.3. AHA!

The AHA! system (De Bra and Calvi, 1998) was developed at the Eindhoven University of Technology. AHA! stands for “Adaptive Hypermedia Architecture”, and provides

adaptive presentation and adaptive navigation support. The architecture of AHA! is presented in Figure 4. The AHA! system uses Java servlets that interact with a combined

domain/adaptation model and with a user mode (De Bra et al., 1999). The adaptive presentation in AHA! is based on inserting/removing fragments techniques. Any

condition on the user model can be used to make a decision of including or excluding a fragment. Whereas the adaptive navigation support in AHA! consists of the following parts:

 The link anchor tag can have a class (style sheet deterring the colour) attribute based on the user model.

 An arbitrary set of icons can be displayed before or after the link anchor, based also on the user model.

 The link destination represents the name of a concept from the domain model.

Figure 4 AHA! architecture (Stash and De Bra, 2004)

2.5.4. My Online Teacher (MOT 1.0)

My Online Teacher (MOT 1.0) is an authoring system for Adaptive Hypermedia and adaptive e-learning systems based on the LAOS framework for authoring of Adaptive Hypermedia (Cristea and de Mooij, 2003a), which it maps onto three levels for authoring:

1. Content level: a hierarchy of domain concepts, with a number of domain attributes; this level also allows for other relatedness relations between concepts. 2. Lesson selection level: also called goal map; filtering and ordering contents at

attribute level or above.

3. Adaptation to student and presentation: this represents the actual adaptation specification.

The Common Adaptation Format (CAF) (Cristea et al., 2007) reflects part of the actual course structure as in the MOT 1.0 database. However, it uses XML representation, which is more suitable for web conversions. CAF instantiates a representation (Cristea et al., 2007) for two of the static modules of the AEH description (1) and (2) from above: domain and lesson maps – or goal maps). Figure 5 illustrates the structure of the CAF XML; x:y defines the range of objects, i.e., the number of objects are between x and y; m stands for many; thus, each CAF file can have one or more domain models, but only one (or no) goal model; each concept in a domain model has only one name.

Figure 5 the Common Adaptation Format (CAF)

The Document Type Definition (DTD) of the CAF XML is shown below:

<?xml version="1.0" encoding="UTF-8"?> <!ELEMENT CAF (domainmodel+, goalmodel?)> <!ELEMENT domainmodel (concept+)>

<!ELEMENT concept (name, attribute*, concept*)> <!ELEMENT attribute (name, contents)>

<!ELEMENT name (#PCDATA)> <!ELEMENT contents (#PCDATA)>

<!ATTLIST link weight CDATA "" label CDATA ""> <!ELEMENT goalmodel (lesson)>

<!ELEMENT lesson (link*, lesson*)>

Thus, a CAF XML file has:

1. Domain model(s) (one or more): containing a set of domain concept(s), each with a set of domain attribute(s) that describe related concept data (the actual

content). An attribute has a name (representing the name of its type), and a content. A concept may have sub-concept(s) that represent associations to other concepts. The creation of the Domain Model, in the context of Adaptive Hypermedia, refers to content organization (Wu, 2002).

2. Goal model: The goal model represents the actual lesson, which may have a set of sub-lesson(s). Each lesson contains a set of link(s). Importantly, each link points to an attribute in the domain model(s), thus reordering the information contained in the initial model(s). Each link has two attributes: weight and label, which are used to determine the adaptation requirements via adaptation strategies (e.g., show links labelled ‘beginner’ to beginner users, and links labelled ‘advanced’ to advanced users).