=Paper=
{{Paper
|id=Vol-3178/CIRCLE_2022_paper_34
|storemode=property
|title=When Truncated Rankings Are Better and How to Measure That – Abstract
|pdfUrl=https://ceur-ws.org/Vol-3178/CIRCLE_2022_paper_34.pdf
|volume=Vol-3178
|authors=Enrique Amigó,Stefano Mizzaro,Damiano Spina
|dblpUrl=https://dblp.org/rec/conf/circle/AmigoMS22
}}
==When Truncated Rankings Are Better and How to Measure That – Abstract==
<pdf width="1500px">https://ceur-ws.org/Vol-3178/CIRCLE_2022_paper_34.pdf</pdf>
<pre>
When Truncated Rankings Are Better and How to
Measure That – Abstract⋆
Enrique Amigó1 , Stefano Mizzaro2 and Damiano Spina3
1
  UNED NLP & IR Group, Madrid, Spain
2
  University of Udine, Italy
3
  RMIT University, Melbourne, Australia


                                         Abstract
                                         In this work we provide both theoretical and experimental contributions for the truncated ranking
                                         evaluation, where systems have a stopping criteria to truncate the ranking at the right position to
                                         avoid retrieving those irrelevant documents at the end. We first define formal properties to analyze
                                         how effectiveness metrics behave when evaluating truncated rankings. Our theoretical analysis shows
                                         that de-facto standard metrics do not satisfy desirable properties to evaluate truncated rankings: only
                                         Observational Information Effectiveness (OIE) – a metric based on Shannon’s information theory –
                                         satisfies them all. We then perform experiments to compare several metrics on nine TREC data sets.
                                         According to our experimental results, the most appropriate metrics for truncated rankings are OIE
                                         and a novel extension of Rank-Biased Precision that adds a user effort factor penalizing the retrieval of
                                         irrelevant documents.

                                         Keywords
                                         Information Retrieval, Evaluation, Evaluation measures, Ranking Cutoff




CIRCLE (Joint Conference of the Information Retrieval Communities in Europe) 2022 is the second joint conference of
the information retrieval communities, July 4-7, 2022, Toulouse, France
⋆
  This work has been published at the SIGIR 2022 main conference.
$ enrique@lsi.uned.es (E. Amigó); mizzaro@uniud.it (S. Mizzaro); mizzaro@uniud.it (D. Spina)
€ https://sites.google.com/view/enriqueamigo/home (E. Amigó); https://www.dimi.uniud.it/mizzaro (S. Mizzaro);
https://www.dimi.uniud.it/mizzaro (D. Spina)
 0000-0003-1482-824X (E. Amigó); 0000-0002-2852-168X (S. Mizzaro); 0000-0001-9913-433X (D. Spina)
                                       © 2022 Copyright 2022 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
    CEUR
    Workshop
    Proceedings
                  http://ceur-ws.org
                  ISSN 1613-0073
                                       CEUR Workshop Proceedings (CEUR-WS.org)

</pre>