Options
2013
Conference Paper
Titel
Comparison of the results of an authorship-based expert recommender against data from a directory of experts
Abstract
Expert recommender systems have been introduced and implemented in various ways. But due to legal issues, the special nature of human-bound knowledge and lack of acceptance by the staff only few case studies and even less evaluations in real world scenarios can be found. This paper presents a real world implementation of a simple expert recommender algorithm based on document authorship as well as an implementation of a directory of experts within one single company. Both have been implemented and therefore it is possible to compare the results of the recommender against data from the expert directory. Thus we wanted to gather first quantitative insights on how many documents a person should be author of before she might feel knowledgeable on a certain topic and even subscribe to it in the directory of experts. These insights will also be presented.