04.11.2014 Views

elektronická verzia publikácie - FIIT STU - Slovenská technická ...

elektronická verzia publikácie - FIIT STU - Slovenská technická ...

elektronická verzia publikácie - FIIT STU - Slovenská technická ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

User Modeling for Personalized Web-Based Systems 219<br />

its updates and a source which contributed the characteristic to the model. Each characteristic<br />

contributes to certain user’s goal with certain relevance.<br />

Because characteristics stored in a model are only estimations of real user features based<br />

on analysis of user’s behavior [6], each characteristic has a confidence, informing about quality<br />

of the estimation.<br />

Complete description of the model can be found at http://nazou.fiit.stuba.sk.<br />

User model in OntoAIMS system. Ontology-based Adaptive Information Management<br />

System (OntoAIMS) provides environment for navigation in and searching for information.<br />

This environment recommends its users (students) the most suitable tasks and helps them<br />

in exploration of the domain [23]. System represents the typical example of employment of<br />

ontologies representing a domain as well as user model. User model, created and maintained<br />

by a specialized component OWL-OLM, is forming a layer on the top of a domain model and<br />

adds user’s perspective to it. This approach is suitable in this case, as the information space<br />

defined in domain model (OS Linux course) is closed and static enough. User model is built<br />

using an interactive dialogue with a user and contains information about each individual<br />

concept like number of concept’s usage, number of correct usages, number of times user<br />

declared concept as understood etc. The model is consequently used to recommend content.<br />

Conclusions on User Model Representation<br />

Ontology is clearly the most promising alternative to represent user model. The basic ide<br />

of ontological modeling is very similar to the way how human think about things in the<br />

world. This allows for inclusion of domain experts into to the process of domain and user<br />

model creation. OWL language is expressive enough to capture even complex relationships<br />

present in reality. Ontologies are the basis for inference mechanisms, which could not only<br />

verify model consistency, but are also able to use conditions, restrictions and rules to infer<br />

new relationships and knowledge. Advantage of ontologies is its shareability, supported by<br />

mapping means provided by OWL.<br />

Disadvantage of ontology based representation is currently the lack of tools and technologies,<br />

(editors, mappers, repositories etc.) in the production-grade quality, which would<br />

provide the same efficiency of work as it is with relational databases and XML-based technologies.<br />

8.2 User Characteristics<br />

When designing adaptive web-based system (and its user modeling part) we need to consider<br />

which features of users will be used for adaptation and thus should be modeled. Five most<br />

popular and useful features are [15]: knowledge, interests, goals and tasks, background and<br />

individual traits.<br />

Knowledge is used mainly in adaptive educational systems, which keep track of user’s<br />

knowledge of the domain as a whole as well as of smaller fragments. Interests (keyword or<br />

concept-based) are becoming important as volume of information space increases and new<br />

information-oriented systems (encyclopedias etc.) are being developed.<br />

Goals and tasks represents the immediate purpose for a user’s work within an adaptive<br />

system [15]. It can be a learning goal or immediate information need.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!