Literatur Details

TitelFeedback Options for a Personal News Recommendation Tool
InCooperation in Classification and Data Analysis, Springer, Berlin, S. 91-98.
Autor(en)Christian Bomhardt and Wolfgang Gaul
KeywordsPersonal Recommendations, Feedback Options, Feedback Models
Abstract Recommendations can help to vanquish the information overload problem on the web. Several web sites provide recommendations for their visitors. If users desire recommendations for sites without this service, they can use browsing agents that give recommendations. In order to obtain user profiles, common agents store interesting and/or uninteresting web pages, use them as training data for the construction of classifiers and give recommendations for unseen web pages. Explicit rating is regarded as most reliable but exhausting method to obtain training data. We present and evaluate alternative feedback options. We show that less exhausting feedback options can be applied successfully.

copyright© Christian Bomhardt

© by InfoTec EDV Consulting & Solutions - Kaiserstuhlstraße 6 - 76275 Ettlingen
Telefon: 07243-94 75 97-0 Fax: 07243-94 75 97-30