Project:

RelFB - Simulated and pseudo relevance feedback

Description:

Experiments on the effectiveness of relevance feedback with real users provide realistic and rich data, but they are time consuming, expensive to perform, and it is impossible to cover extensively all possible test cases. This makes simulation for rapid  testing desirable. Project RelFB aims at developing a simulation methodology to explore the limits of the effectiveness of relevance feedback.  Currently, we focus on simulating specifically the user viewpoint via utilizing user modeling together with cumulated-gain based evaluation methods.

Recent publications:
http://liinwww.ira.uka.de/csbib?query=au:KeskustaloH*&maxnum=200&sort=year

Duration:

2005-2008

Researchers:

Mr. Heikki Keskustalo, Acad. Prof. Kalervo Järvelin, Dr. Ari Pirkola

Updated:

3 April 2007. Responsibility for updating: HK.