Filters








4,887 Hits in 6.3 sec

Applying the Query Change Retrieval Model on Session Search-Georgetown at TREC 2013 Session Track

Sicong Zhang, Hui Yang
2013 Text Retrieval Conference  
Comparing the retrieval accuracy in RL1 with that in RL2 and RL3, we obtain 46.1% and 46.6% improvements, respectively.  ...  In TREC 2013 Session Track, we experiment the Query Change Model (QCM) for session search on the new document collection ClueWeb12 CatB.  ...  Table 4 . nDCG@10 scores for TREC2013 runs. 4 (RL2 improvements with RL1, RL3 improvements compared with RL2) GUrun1 Improvement GUrun2 Improvement GUrun3 Improvement RL1 0.1043 - 0.1013 - 0.105 - RL2  ... 
dblp:conf/trec/ZhangY13 fatcat:q7adgrsor5h6ximr5kf7yxkd74

University of Glasgow at TREC 2011: Experiments with Terrier in Crowdsourcing, Microblog, and Web Tracks

Richard McCreadie, Craig Macdonald, Rodrygo L. T. Santos, Iadh Ounis
2011 Text Retrieval Conference  
Meanwhile, we continue to build upon our novel xQuAD framework and data-driven ranking approaches within Terrier to achieve effective and efficient ranking for the TREC Web track.  ...  In TREC 2011, we focus on tackling the new challenges proposed by the pilot Crowdsourcing and Microblog tracks, using our Terrier Information Retrieval Platform.  ...  ACKNOWLEDGEMENTS We would like to thank Amazon's Mechanical Turk for funding support provided for experiments detailed in this paper.  ... 
dblp:conf/trec/McCreadieMSO11 fatcat:gdi4zp7ozvbmdmur7h6tznuri4

Effective Structured Query Formulation for Session Search

Dongyi Guan, Hui Yang, Nazli Goharian
2012 Text Retrieval Conference  
The strict form of query formulation achieves an improvement of 13.5% and the relaxed form achieves an improvement of 17.8% on nDCG@10 on TREC 2011 query sets.  ...  We find that previous vs. current achieves the best search accuracy. For retrieval, we first retrieve a large pool of documents for q n .  ...  AP for TREC 2012 runs.  ... 
dblp:conf/trec/GuanYG12 fatcat:7qaufwczynczhgaqht73q4orwi

Evaluating high accuracy retrieval techniques

Chirag Shah, W. Bruce Croft
2004 Proceedings of the 27th annual international conference on Research and development in information retrieval - SIGIR '04  
The experiments done on TREC data provide support for the approach of using MRR and incorporating QA techniques for getting high accuracy in ad-hoc retrieval task.  ...  We evaluate three different methods for modifying queries to achieve high accuracy.  ...  System Dataset Number of docs MRR Correct answers (median) (median) TREC-8 TREC Disks 4, 5 528,000 For 50-byte responses: 0.2610 39.50% (1999) For 250-byte response: 0.3830 52.50% TREC-9  ... 
doi:10.1145/1008992.1008996 dblp:conf/sigir/ShahC04 fatcat:77gxtci5nbe3xitcitvkc5bmpi

On bidirectional English-Arabic search

M. Aljlayl, O. Frieder, D. Grossman
2002 Journal of the American Society for Information Science and Technology  
We found that query expansion after translation via posttranslation approach yields significant improvement on the performance of the English-Arabic CLIR.  ...  We also empirically evaluate the effectiveness of the Arabic-English MT-based approach using short, medium, and long queries of TREC topics and collections.  ...  time performance) of AIRE search engine to run the translated titles, descriptions and narratives fields of topics TREC-7 and TREC-9.Table 21.  ... 
doi:10.1002/asi.10143 fatcat:dzqja4wqijelncpqw6ya3k2goe

Clairvoyance Corporation Experiments in the TREC 2003 High Accuracy Retrieval from Douments (HARD) Track

James G. Shanahan, Jeffrey Bennett, David A. Evans, David A. Hull, Jesse Montgomery
2003 Text Retrieval Conference  
Figure 4 presents a comparison of the TREC Median and Maximum to our submitted baseline run and our best submitted run (CLAI1NG).  ...  And if the user runs out of time or simply fails to judge a group, the default value for a group is "Unjudged".  ... 
dblp:conf/trec/ShanahanBEHM03 fatcat:hmif6mbbe5c7jopbrdzeirxriy

A Case for Automatic System Evaluation [chapter]

Claudia Hauff, Djoerd Hiemstra, Leif Azzopardi, Franciska de Jong
2010 Lecture Notes in Computer Science  
These findings support the case for automatic system evaluation and motivate further research.  ...  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.  ...  strategy where a subset of topics is built by each time adding the topic with the highest estimation accuracy (single topic performance).  ... 
doi:10.1007/978-3-642-12275-0_16 fatcat:gozggcd5qbgjhnyhgugk4nzrha

The Role of Anchor Text in ClueWeb09 Retrieval

Vo Ngoc Anh, Alistair Moffat
2010 Text Retrieval Conference  
This report describes the work done at The University of Melbourne with the ClueWeb09 data corpus for the Web Track of TREC-2009 and TREC-2010, and for the Session Track of TREC-2010.  ...  We found that the impact-based retrieval model works well for the corpus, and that, along with some other factors, the use of an anchor text collection significantly boosts the retrieval effectiveness.  ...  We thank RMIT University (Australia) for letting us to use their high-performance computing cluster for all the experiments described in this report.  ... 
dblp:conf/trec/AnhM10 fatcat:dfg6v3rnhzeslkuvluax7iefl4

Ranking Retrieval Systems without Relevance Assessments: Revisited

Tetsuya Sakai, Chin-Yew Lin
2010 NTCIR Conference on Evaluation of Information Access Technologies  
Our extensive experiments using graded-relevance test collections from TREC and NTCIR compare several existing methods for ranking systems without relevance assessments.  ...  We show that (a) The simplest method of forming "pseudo-qrels" based on how many systems returned each pooled document performs as well as any other existing method; and that (b) the NTCIR system rankings  ...  We thank Ellen Voorhees for checking for us that the TREC 2004 robust track relevance assessments were done after sorting by document IDs.  ... 
dblp:conf/ntcir/SakaiL10 fatcat:gpppeiguczgwljs7ad7riqlb6e

Optimization in CLARIT TREC-8 Adaptive Filtering

ChengXiang Zhai, Peter Jansen, Norbert Roma, Emilia Stoica, David A. Evans
1999 Text Retrieval Conference  
In this paper, we describe the system and methods used for the CLARITECH entries in the TREC-8 Filtering Track.  ...  Our results show that profile-specific frequent updating indeed improves filtering performance.  ...  The second hypothesis we tested in TREC-8 is that we can improve the filtering performance by improving individual filtering system components independently.  ... 
dblp:conf/trec/ZhaiJRSE99 fatcat:lg73iybxyrhojh3wl427zfkmvi

Methods for ranking information retrieval systems without relevance judgments

Shengli Wu, Fabio Crestani
2003 Proceedings of the 2003 ACM symposium on Applied computing - SAC '03  
precision, R-precision, and precision and different document levels.  ...  An extensive experimentation was conducted to evaluate the effectiveness of the proposed methods using various different standards Information Retrieval evaluation measures for the ranking, like average  ...  The experiment is conducted for all 5 TREC collections, and each collection has runs 20 times for 50 queries.  ... 
doi:10.1145/952686.952693 fatcat:k2g55wzdaze3vprlkkj2ovozsu

Methods for ranking information retrieval systems without relevance judgments

Shengli Wu, Fabio Crestani
2003 Proceedings of the 2003 ACM symposium on Applied computing - SAC '03  
precision, R-precision, and precision and different document levels.  ...  An extensive experimentation was conducted to evaluate the effectiveness of the proposed methods using various different standards Information Retrieval evaluation measures for the ranking, like average  ...  The experiment is conducted for all 5 TREC collections, and each collection has runs 20 times for 50 queries.  ... 
doi:10.1145/952532.952693 dblp:conf/sac/WuC03 fatcat:vbu3fpdv45avvlqqywd2f2e42a

Seven Hypothesis about Spam Filtering

William S. Yerazunis
2006 Text Retrieval Conference  
For TREC 2006 , the CRM114 team considered several different hypothesis on the topic of spam filtering.  ...  We also found a fascinating countercorrellation between 1ROCA% and the final accuracy of a filter (the accuracy of the filter for the final 10% of the corpus).  ...  For the third CRM114 run, we considered the TREC 2005 curve set and noted that the CRM114 OSB classifier exhibited a strange dip in the curve, where the accuracy became very good, and then seemed to deteriorate  ... 
dblp:conf/trec/Yerazunis06 fatcat:gw4nvo46vrdqtfp3v7fvvrypvm

OSC at TREC 2020 - News track's Background Linking Task

Nathan Day, Dan Worley, Tim Allison
2020 Text Retrieval Conference  
OSC's NER and document embedding runs both produced better results than MLT with default parameters.  ...  OSC submitted four runs: MLT with default parameters, MLT with tuned parameters, MLT with a new named entity recognition (NER) field, and MLT re-scored by document embedding similarity.  ...  Acknowledgments Acknowledgments The authors would like to thank Samantha Toet and Charlie Hull for their editorial feedback.  ... 
dblp:conf/trec/DayWA20 fatcat:4tyweacmubfiphdarht6m7yuda

Predicting the effectiveness of queries and retrieval systems

Claudia Hauff
2010 SIGIR Forum  
Smaller improvements up to τ = 0.1 are also achieved for TREC Vol. 4+5 and MaxVAR, where τ peaks at µ = 2.5 × 10 4 .  ...  SQE Experiment Given the 1000 generated rankings per correlation coefficient interval c i and the 500 run pairs (run base /run aqe ) for each setting of θ and m, SQE is thus performed 500000 times for  ...  (TREC Vol. 4+5) [148] , WT10g [132] and GOV2 [38] .  ... 
doi:10.1145/1842890.1842906 fatcat:jkesk5hrvfe77bg5xr7yg7bbie
« Previous Showing results 1 — 15 out of 4,887 results