A Case for Automatic System Evaluation
Ranking a set retrieval systems according to their retrieval effectiveness without relying on relevance judgments was first explored by Soboroff et al. . Over the years, a number of alternative approaches have been proposed, all of which have been evaluated on early TREC test collections. In this work, we perform a wider analysis of system ranking estimation methods on sixteen TREC data sets which cover more tasks and corpora than previously. Our analysis reveals that the performance of system ranking estimation approaches varies across topics. This observation motivates the hypothesis that the performance of such methods can be improved by selecting the “right” subset of topics from a topic set. We show that using topic subsets improves the performance of automatic system ranking methods by 26% on average, with a maximum of 60%. We also observe that the commonly experienced problem of underestimating the performance of the best systems is data set dependent and not inherent to system ranking estimation. These findings support the case for automatic system evaluation and motivate further research.
Unable to display preview. Download preview PDF.
- 1.Rank Correlation Methods. Hafner Publishing Co., New York (1955)Google Scholar
- 2.Amitay, E., Carmel, D., Lempel, R., Soffer, A.: Scaling ir-system evaluation using term relevance sets. In: SIGIR 2004, pp. 10–17 (2004)Google Scholar
- 4.Aslam, J.A., Pavlu, V., Yilmaz, E.: A statistical method for system evaluation using incomplete judgments. In: SIGIR 2006, pp. 541–548 (2006)Google Scholar
- 5.Aslam, J.A., Savell, R.: On the effectiveness of evaluating retrieval systems in the absence of relevance judgments. In: SIGIR 2003, pp. 361–362 (2003)Google Scholar
- 6.Carterette, B., Allan, J.: Incremental test collections. In: CIKM 2005, pp. 680–687 (2005)Google Scholar
- 7.Diaz, F.: Performance prediction using spatial autocorrelation. In: SIGIR 2007, pp. 583–590 (2007)Google Scholar
- 8.Efron, M.: Using multiple query aspects to build test collections without human relevance judgments. In: ECIR 2009, pp. 276–287 (2009)Google Scholar
- 9.Guiver, J., Mizzaro, S., Robertson, S.: A few good topics: Experiments in topic set reduction for retrieval evaluation. To appear in TOISGoogle Scholar
- 10.Krovetz, R.: Viewing morphology as an inference process. In: SIGIR 1993, pp. 191–202 (1993)Google Scholar
- 11.Mizzaro, S., Robertson, S.: Hits hits trec: exploring ir evaluation results with network analysis. In: SIGIR 2007, pp. 479–486 (2007)Google Scholar
- 13.Soboroff, I., Nicholas, C., Cahan, P.: Ranking retrieval systems without relevance judgments. In: SIGIR 2001, pp. 66–73 (2001)Google Scholar
- 16.Wu, S., Crestani, F.: Methods for ranking information retrieval systems without relevance judgments. In: Matsui, M., Zuccherato, R.J. (eds.) SAC 2003. LNCS, vol. 3006, pp. 811–816. Springer, Heidelberg (2004)Google Scholar