LogCLEF 2009: The CLEF 2009 multilingual Logfile analysis track overview

Abstract

Log data constitute a relevant aspect in the evaluation process of the quality of a search engine and the quality of a multilingual search service; log data can be used to study the usage of a search engine, and to better adapt it to the objectives the users were expecting to reach. The interest in multilingual log analysis was promoted by the Cross Language Evaluation Forum (CLEF) for the first time with a track named LogCLEF. LogCLEF is an evaluation initiative for the analysis of queries and other logged activities as expression of user behavior. The goal is the analysis and classification of queries in order to understand search behavior especially in multilingual contexts and ultimately to improve search systems. Two tasks were defined: Log Analysis and Geographic Query Identification (LAGI) which aimed at the identification of queries for geographic content and Log Analysis for Digital Societies (LADS) which was based on analyzing the user behavior of the search logs the service of The European Library. Five groups using a variety of approaches submitted experiments. The data for the track, the evaluation methodology and results are presented and discussed. © 2010 Springer-Verlag Berlin Heidelberg.