In 2009, all computations were done with SOIRE, a software for IR evaluation based on a service-oriented architecture and double-checked against trec_eval, the standard program for evaluation used in the TREC evaluation campaign.
in 2010 we have used the trec_eval version 9.0 for both Prior Art and Classification tasks. To compute the PRES score, we have used the script provided by the authors of this measure.
For each experiment submitted to the Prior Art types of tasks we computed 10 standard IR measures:
- Precision, Precision@5, Precision@10, Precision@100
- Recall, Recall@5, Recall@10, Recall@100
- nDCG(with reduction factor given by a logarithm in base 10)
- PRES (starting 2010)
For the Classification task we computed the following measures:
- Precision@1, Precision@5, Precision@10, Precision@25, Precision@50
- Recall@5, Recall@25, Recall@50
- F1 at 5, 25, and 50
All evaluation results are published as IRF technical reports.
- IRF-TR-2009-00001: Florina Piroi, Giovanna Roda, Veronika Zenz, CLEF-IP 2009 Evaluation Summary, July 2009, download PDF
- IRF-TR-2010-00002: Florina Piroi, Giovanna Roda, Veronika Zenz, CLEF-IP 2009: Evaluation Summary - Part Two, July 2010, download PDF