Skip to content

Runs - Million Query 2007

exegyexact

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: exegyexact
  • Participant: exegy.indeck
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: The results were generated automatically using the Exegy TextMiner engine. The dataset is not indexed, the engine performs search on the data as it streams through. The engine looks for the queries exactly as they appear in the query file. Documents are ranked according to the number of occurrences of a particular query in a document.

ffind07c

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: ffind07c
  • Participant: ualaska.newby
  • Track: Million Query
  • Year: 2007
  • Submission: 6/17/2007
  • Type: automatic
  • Task: official
  • Run description: I used the TREC TB track qrels to choose a subset of GOV2 to search. This is a distributed/grid IR simulation.

ffind07d

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: ffind07d
  • Participant: ualaska.newby
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: This is a large-scale distributed/grid IR simulation. I used the qrels from the previous TREC Terabyte tracks to pick a subset of GOV2 to search.

hedge0

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: hedge0
  • Participant: northeasteru.aslam
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: We used several standard Lemur built in systems (tfidf_bm25, tfidf_log, kl_abs,kl_dir,inquery,cos, okapi) and combined their output (metasearch) using the hedge algorithm.

hitir2007mq

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: hitir2007mq
  • Participant: heilongjiang-it.qi
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: We are a new group in IR society from Heilongjiang Institute of Technology, China. This is our first time to participate TREC evaluation. LEMUR is used in our run.

indriDM

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: indriDM
  • Participant: umass.allan
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: we use don's dependency model and terms in topic title to generate queries, and run by using indri

indriDMCSC

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: indriDMCSC
  • Participant: umass.allan
  • Track: Million Query
  • Year: 2007
  • Submission: 6/16/2007
  • Type: automatic
  • Task: official
  • Run description: use all the terms in the topic title and don's denpendence model, but we do spell checking by using the typical aspell(or ispell) tool in unix system, and then use wsyn operator of indri to put the spell checking results into the queries

indriQL

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: indriQL
  • Participant: umass.allan
  • Track: Million Query
  • Year: 2007
  • Submission: 6/16/2007
  • Type: automatic
  • Task: official
  • Run description: use all the terms in the topic title, and directly use combination operator of indri

indriQLSC

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: indriQLSC
  • Participant: umass.allan
  • Track: Million Query
  • Year: 2007
  • Submission: 6/16/2007
  • Type: automatic
  • Task: official
  • Run description: use all the terms in the topic title but we do spell checking by using the typical aspell(or ispell) tool in unix system, and then use weight and combination operator of indri to run the query

JuruSynE

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: JuruSynE
  • Participant: ibm.carmel
  • Track: Million Query
  • Year: 2007
  • Submission: 6/16/2007
  • Type: automatic
  • Task: official
  • Run description: Basic Juru run. Docs are scored according to their textual similarity to the query and their number of in-links. Queries are expanded by a short list of sysnoyms related to the gov domain

LucSpel0

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: LucSpel0
  • Participant: ibm.carmel
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: Lucene, doc text + anchor text, text queries with phrase and proximity elements, stopwords, query expansion by synonyms and spell correction (index based), modified doc length normalization, modified tf()

LucSyn0

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: LucSyn0
  • Participant: ibm.carmel
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: Lucene, doc text + anchor text, text queries with phrase and proximity elements, stopwords, synonyms query expansion, modified doc length normalization, modified tf().

LucSynEx

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: LucSynEx
  • Participant: ibm.carmel
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: Lucene, doc text + anchor text, text queries with phrase and proximity elements, stopwords, synonyms query expansion (expansions will have greater impact in this run, comparing to LucSyn0) , modified doc length normalization, modified tf().

rmitbase

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: rmitbase
  • Participant: rmitu.scholer
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: Zettair Dirichlet smoothed language model run.

sabmq07a1

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: sabmq07a1
  • Participant: sabir.buckley
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: Standard smart ltu.Lnu run

sabmq07sam

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: sabmq07sam
  • Participant: sabir.buckley
  • Track: Million Query
  • Year: 2007
  • Submission: 5/25/2007
  • Type: automatic
  • Task: trial
  • Run description: Straight simple Lnu-ltu weighted vector run

UAmsT06tAnLM

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UAmsT06tAnLM
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 5/23/2007
  • Type: automatic
  • Task: trial
  • Run description: Anchor-texts index, using the Snowball stemming algorithm, standard multinomial language model with Jelinek-Mercer smoothing, lambda = .9

UAmsT06tAnVS

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UAmsT06tAnVS
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 5/23/2007
  • Type: automatic
  • Task: trial
  • Run description: Anchor-texts index, using the Snowball stemming algorithm, standard Lucene vector-space model

UAmsT06tTeLM

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UAmsT06tTeLM
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 5/23/2007
  • Type: automatic
  • Task: trial
  • Run description: Full-text index, using the Snowball stemming algorithm, standard multinomial language model with Jelinek-Mercer smoothing, lambda = .9

UAmsT06tTeVS

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UAmsT06tTeVS
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 5/23/2007
  • Type: automatic
  • Task: trial
  • Run description: Full-text index, using the Snowball stemming algorithm, standard Lucene vector-space model

UAmsT06tTiLM

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UAmsT06tTiLM
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 5/23/2007
  • Type: automatic
  • Task: trial
  • Run description: Title fields index, using the Snowball stemming algorithm, standard multinomial language model with Jelinek-Mercer smoothing, lambda = .9

UAmsT07MAnLM

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: UAmsT07MAnLM
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: Anchor-texts index, using the Snowball stemming algorithm, standard multinomial language model with Jelinek-Mercer smoothing, lambda = .9

UAmsT07MSm8L

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UAmsT07MSm8L
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 7/16/2007
  • Type: automatic
  • Task: unpooled
  • Run description: Weighted CombSUM of language model runs (lambda = .9) on the full-text index (relative weight 0.8), anchor-text index (relative weight 0.1), and titles index (relative weight 0.1), all using the Snowball stemming algorithm.

UAmsT07MSum6

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: UAmsT07MSum6
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: Weighted CombSUM of vector-space model runs on the full-text index (relative weight 0.6), anchor-text index (relative weight 0.2), and titles index (relative weight 0.2), all using the Snowball stemming algorithm.

UAmsT07MSum8

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: UAmsT07MSum8
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: Weighted CombSUM of vector-space model runs on the full-text index (relative weight 0.8), anchor-text index (relative weight 0.1), and titles index (relative weight 0.1), all using the Snowball stemming algorithm.

UAmsT07MTeLM

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UAmsT07MTeLM
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 7/16/2007
  • Type: automatic
  • Task: unpooled
  • Run description: Full-text index, using the Snowball stemming algorithm, standard multinomial language model with Jelinek-Mercer smoothing, lambda = .9

UAmsT07MTeVS

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: UAmsT07MTeVS
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: Full-text index, using the Snowball stemming algorithm, standard Lucene vector-space model.

UAmsT07MTiLM

Results | Participants | Proceedings | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: UAmsT07MTiLM
  • Participant: uamsterdam.deRijke
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: Title fields index, using the Snowball stemming algorithm, standard multinomial language model with Jelinek-Mercer smoothing, lambda = .9

UiucMQbl

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UiucMQbl
  • Participant: uc-zhai
  • Track: Million Query
  • Year: 2007
  • Submission: 7/6/2007
  • Type: automatic
  • Task: unpooled
  • Run description: baseline run, using axiomatic approach,

UiucMQqe1

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UiucMQqe1
  • Participant: uc-zhai
  • Track: Million Query
  • Year: 2007
  • Submission: 7/9/2007
  • Type: automatic
  • Task: unpooled
  • Run description: using axiomatic approach + semantic query expansion. We used the top 100 snippets returned by Yahoo as resources to select expanded query terms.

UiucMQqe2

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP)

  • Run ID: UiucMQqe2
  • Participant: uc-zhai
  • Track: Million Query
  • Year: 2007
  • Submission: 7/9/2007
  • Type: automatic
  • Task: unpooled
  • Run description: using axiomatic approach + semantic query expansion. We used the top 100 snippets returned by Yahoo as resources to select expanded query terms.

umelbexp

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: umelbexp
  • Participant: umelbourne.ngoc-ahn
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: Submit query to public web search engine, retrieve snippet information for top 5 documents, add unique terms from snippets to query, run expanded query using same similarity metric as umelbstd

umelbimp

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: umelbimp
  • Participant: umelbourne.ngoc-ahn
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: standard impact-based ranking

umelbsim

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: umelbsim
  • Participant: umelbourne.ngoc-ahn
  • Track: Million Query
  • Year: 2007
  • Submission: 6/19/2007
  • Type: automatic
  • Task: official
  • Run description: merging of the language modelling and the impact runs

umelbstd

Results | Participants | Input | Summary (tb-topics) | Summary (statMAP) | Appendix

  • Run ID: umelbstd
  • Participant: umelbourne.ngoc-ahn
  • Track: Million Query
  • Year: 2007
  • Submission: 6/18/2007
  • Type: automatic
  • Task: official
  • Run description: topic-only run using a similarity metric based on a language model with Dirichlet smoothing as describe by Zhai and Lafferty (2004)