Method for User Preferences Acquisition (tool UPreA)
Finds initial user preference in form of fuzzy sets and aggregation function.
Institution: | Pavol Jozef Šafárik University |
Technologies used: | Java, Sesame |
Inputs: | Domain ontology, User Ontology, user inputs |
Outputs: | User preferences (fuzzy sets) |
Documentation: | HTML, doc, JavaDoc |
Distribution packages: | zip |
Addressed Problems
Methods for finding best objects usually use some representation of user preferences. They can be represented as sets (or intervals) of prefered attribute values. Instead of classical crisp sets, we use fuzzy sets with membership function range [0,1].
The problem is how to acquire user preferences effectively and correctly. If we let user explain his preferences in detail, it may be boring and uninteresting. On the other hand, if we ask just a few questions, we get only vague idea what his preferences might be. This problem is adresses by UPreA tool. It provides user with initial fuzzy sets which can be used in the process of finding suitable job offers and further modified.
Description
We consider three different approaches to acquire user preferences.
- Direct input from user. This approach uses graphic interface shown on the following figure. The interface can be a part of registration form.
- Filling fuzzy sets according to other similar users. Fuzzy sets are of four basic types: ascending, descending, hill and valley. These types indicate the shape of fuzzy membership function. Users belong to some user group according to their personal attributes like gender, education, etc. Therefore we are able to remember prevalent type of fuzzy set for every group of similar users. If a new user does not input his preferences, we find a group he belongs to and we assign him fuzzy sets of prevalent type for every attribute.
- Using defaults. This approach is used when there is no other user in the same group. Default fuzzy sets are the most frequented types, e.g. ascending type for salary.

Graphic interface inside applet for manual input of fuzzy sets.

Index of user preferences.
Acquired fuzzy sets and weights are provided to other tools. IGAP and Top-k tools recently use them in our integrated chain.
References
- Vaneková, V.: Metódy získavania používateľských preferencií In: ZNALOSTI 2008: zborník príspevkov. Vydavateľstvo STU, 2008. ISBN 978-80-227-2827-0, pages 387-390.
- Gurský, P., Horváth, T., Jirásek, J., Krajči, S., Novotný, R., Vaneková, V., Vojtáš, P.: Knowledge Processing for Web Search – An Integrated Model. In: Studies in Computational Intelligence (vol. 78), Springer, 2007, ISSN 1860-949X, ISBN 978-3-540-74929-5, pages 95-104.
- Gurský, P., Horváth, T., Jirásek, J., Krajči, S., Novotný, R., Vaneková, V., Vojtáš, P.: Web Search with Variable User Model. In: DATAKON 2007, Sborník databázové konference, ISBN 978-80-7355-076-9, pages 111-121.
- Vaneková, V.: Reprezentácia a spracovanie používateľských preferencií v RDF. In: Proceedings of MIS 2007. ISBN 978-80-7378-033-3, pages 98-107.
- Gurský, P., Horváth, T., Novotný, R., Vaneková, V., Vojtáš, P.:UPRE: User preference based search system. In: Proceedings of 2006 IEEE/WIC/ACM International Conference on Web Intelligence (WI 2006), ISBN 0-7695-2747-7, pages 841-844.