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 225<br />

Personis user model<br />

server<br />

View N<br />

Common<br />

scrutability<br />

interface<br />

View 1<br />

View 2<br />

AH N<br />

...<br />

Scrutability<br />

interface<br />

AH 1<br />

Scrutability<br />

interface<br />

AH 2<br />

Scrutability<br />

interface<br />

User<br />

Figure 8-4. Integration of Personis server with adaptive systems, according to [37].<br />

SemanticLog<br />

SemanticLog [5] is a user modeling server based on OWL ontologies. It accepts notifications<br />

about occurrence of semantically defined events from multiple adaptive applications<br />

(informing about an action performed by an individual user) and stores these events as an evidence.<br />

Subsequently, it notifies registered user modeling inference agents about new event<br />

to process. After its processing, inference agents may update the user model in a central<br />

repository, where it is available to the adaptive applications.<br />

The idea of SemanticLog is similar to CUMULATE. However, the representation of events<br />

in the log using semantics based on ontologies allows for better separation (and thus better<br />

reusability) of the log processing from the presentation tools and the server itself (responsible<br />

mainly for log acquisition and storage). Inference agents work with semantics of the events<br />

defined in the shared event ontology and do not need to know about the applications which<br />

supplied the events [6].<br />

8.5 Sources for User Modeling<br />

Disadvantage of an overlay user model representation is its complicated initialization. When<br />

user logs into the system for the first time, the system does not have enough information<br />

to provide efficient personalization (if it does not have any access to user modeling server<br />

holding information about the new user). This lack of new user characteristics is called

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

Saved successfully!

Ooh no, something went wrong!