No abstract

No abstract

Documents are recommended by computer-based systems normally according to their prominence in the document reference network. Based on the requirements identified in a concrete use case for recommending scientific publications, the paper claims that merely measuring prominence is insufficient for high quality recommendations. We propose to use information from a trust network in addition to the document network in order to improve and to personalize recommendations. A trust-enhanced visibility measure integrates trust information and the classical reference based measures. A simulation study applies the new visibility measure to the presented use case.

The considerable high quality of Wikipedia articles is often accredited to the large number of users who contribute to Wikipedia's encyclopedia articles, who watch articles and correct errors immediately. In this paper, we are in particular interested in a certain type of Wikipedia articles, namely, the featured articles -articles marked by a community's vote as being of outstanding quality. The German Wikipedia has the nice property that it has two types of featured articles: excellent and worth reading. We explore on the German Wikipedia whether only the mere number of contributors makes the difference or whether the high quality of featured articles results from having experienced authors contributing with a reputation for high quality contributions. Our results indicate that it does matter who contributes.

No abstract