Skip to content

Runs - Enterprise 2008

4FvfI

Results | Participants | Proceedings | Input | Summary

  • Run ID: 4FvfI
  • Participant: inria
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: aecae434a6eb8508dcc7c6fc2d56a464

8T0eZ

Results | Participants | Proceedings | Input | Summary

  • Run ID: 8T0eZ
  • Participant: st.petersburg
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: 7f3593e22dade848d95e33adfbc35178

DERIrun1

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun1
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/23/2008
  • Type: automatic
  • Task: expert
  • MD5: ab7ccf6e66587f7ce501e743484e1197
  • Run description: Profiling approach. Concatenate documents for an expert profile. Rank using an learned ranking function on 2007 data and topics.

DERIrun2

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun2
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/24/2008
  • Type: automatic
  • Task: expert
  • MD5: 84dade757a652608b046adee55723a5e
  • Run description: Query Only. Profiling approach. Concatenate documents for an expert profile. Rank using an learned ranking (27) function on 2007 data and topics.

DERIrun3

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun3
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/24/2008
  • Type: automatic
  • Task: expert
  • MD5: 14f3ee77b3f920d333090b726c55231f
  • Run description: Two-Stage approach. Ranked documents using evolved scheme (es7) and summed top 5 documents relating to an expert. Query fields.

DERIrun4

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun4
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/24/2008
  • Type: automatic
  • Task: expert
  • MD5: 1d21b31f7fc83a195f617a693026d5b3
  • Run description: Two-Stage approach. Ranked documents using evolved scheme (es) and summed top 5 documents relating to an expert. Query & Narr fields.

DERIrun5

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun5
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/23/2008
  • Type: automatic
  • Task: document
  • MD5: 11dbd6f548203c40f6880b516be342a4
  • Run description: Used Query field and a novel learned term-weighting scheme.

DERIrun6

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun6
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/23/2008
  • Type: automatic
  • Task: document
  • MD5: 9db1e9e4f6fd2164a19472977c0b801f
  • Run description: Query Field only. Used a novel learned term-weighting scheme for ranking (es7).

DERIrun7

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun7
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/23/2008
  • Type: automatic
  • Task: document
  • MD5: b91ab82950aab9a4b3afeef9c73def1d
  • Run description: Used Query and Narrative fields and an novel learned term-weighting scheme (es).

DERIrun8

Results | Participants | Proceedings | Input | Summary

  • Run ID: DERIrun8
  • Participant: DERI_IR_GROUP
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/23/2008
  • Type: automatic
  • Task: document
  • MD5: ee83326df7adb01e330be97b752e1c25
  • Run description: Query and narrative fields used. Ranked using novel learned term-weighting scheme (es7).

FDUBase

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDUBase
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/2/2008
  • Type: automatic
  • Task: document
  • MD5: 56a917b165fe0cbcb6a918ecac569c4e
  • Run description: We use Lemur as the search engine for this run and we do not do extra process.

FDUEmail

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDUEmail
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/2/2008
  • Type: automatic
  • Task: document
  • MD5: 1bfa90000363c42704340db58cfd7b51
  • Run description: Based on the basic run, we give higher score to the pages which have email.

FDUExpand

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDUExpand
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/2/2008
  • Type: automatic
  • Task: document
  • MD5: 67021c22c3ca6a9a3e34424c577f9bc2
  • Run description: We use word net to make query expansion and use lemur as the search engine.

FDUExpBase

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDUExpBase
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/9/2008
  • Type: automatic
  • Task: expert
  • MD5: 4f0d3a3c77b32df5bfcba938069b5fb9
  • Run description: This run used the classical expert search model. Given a topic, we search the files related to the topic and then search the experts related to the files.

FDUExpRes

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDUExpRes
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/9/2008
  • Type: automatic
  • Task: expert
  • MD5: a21b0d464e546a5107e4bc51ae938564
  • Run description: As the resume pages always give more information of the experts than other web pages, we extract resume pages from the corpus and regard them as important evidence for the experts' authority to improve the basic run.

FDUExpRole

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDUExpRole
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/9/2008
  • Type: automatic
  • Task: expert
  • MD5: 7e20a95327cd65d087fd9a2c78822f1e
  • Run description: Based on the basic run, we give the candidate more scores if the context of the candidate's name show the role of the candidate is probably an expert.

FDURoleRes

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDURoleRes
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/9/2008
  • Type: automatic
  • Task: expert
  • MD5: 876ffe1c25ba3d2ba3ba44a1290f7d86
  • Run description: We combine the factors of the context of the candidates' names and the resume pages to improve the basic run.

FDUUrl

Results | Participants | Proceedings | Input | Summary

  • Run ID: FDUUrl
  • Participant: wim-lab.fudan
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/2/2008
  • Type: automatic
  • Task: document
  • MD5: 491584b95eebedb2f3ed75d490b4b173
  • Run description: We define several simple rules to select informative pages by analyzing url. Based on the basic run, we give those pages higher scores.

ICTI3Sdoc01

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sdoc01
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 2a03f94d145a3d8ad524ad0b862135c6
  • Run description: As to this run, we employed BM25F text retrieval model, and only "query" field is used.

ICTI3Sdoc02

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sdoc02
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 401d286e16dbb10c164e0c9d6b9e9254
  • Run description: As to this run, we employed Language Model text retrieval, and only the "query" field is used.

ICTI3Sdoc03

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sdoc03
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 30487ef8a3e604730d9b181fc9c7f6db
  • Run description: As to this run, we employ the BM25F text retrieval model and only the "query" field is used. The query expansion technique is adopted.

ICTI3Sdoc04

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sdoc04
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: bae5e73f3917adf7a69bc154b9c75cef
  • Run description: As to this run, we employ the language model and only the "query" field is used. The query expansion technique is adopted.

ICTI3Sexp01

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sexp01
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/7/2008
  • Type: automatic
  • Task: expert
  • MD5: 05309d38d583f38b8fec2a581b9ff37e
  • Run description: Text Retrieval Model BM25; Expert Search Model VOTING.

ICTI3Sexp02

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sexp02
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/7/2008
  • Type: automatic
  • Task: expert
  • MD5: 1483957b25026f240d5716e508ca17a7
  • Run description: Text Retrieval Modal BM25F; Expert Search Modal VOTING; KL is employed in Query Extension.

ICTI3Sexp03

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sexp03
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/7/2008
  • Type: automatic
  • Task: expert
  • MD5: 1f678592a6a4e81d03bce74b4751623f
  • Run description: Two-stage language model is employed.

ICTI3Sexp04

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTI3Sexp04
  • Participant: I3S_Group_of_ICT
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/7/2008
  • Type: automatic
  • Task: expert
  • MD5: 4dbf030f599d2f30fa16fd8f4458f5ec
  • Run description: As to this run, we construct a description document for each candidate expert. Then, the traditional text retrieval technique is employed to retrieve the desired experts.

Krcy7

Results | Participants | Proceedings | Input | Summary

  • Run ID: Krcy7
  • Participant: st.petersburg
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: 288c5837042ebc6909a2dcf4fd89c5ee

LiaExp08

Results | Participants | Proceedings | Input | Summary

  • Run ID: LiaExp08
  • Participant: lia-talne
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: automatic
  • Task: expert
  • MD5: ad98551a0371f42c1602f0971b666311
  • Run description: Indri queries on corpus of e-mail addresses. One file per e-mail.

LiaIcAuto

Results | Participants | Proceedings | Input | Summary

  • Run ID: LiaIcAuto
  • Participant: lia-talne
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: f81d0afcf7477bfdadf2039c826d601a
  • Run description: Automatic run using SOM document similarity with narratives

LiaIcExp08

Results | Participants | Proceedings | Input | Summary

  • Run ID: LiaIcExp08
  • Participant: lia-talne
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: manual
  • Task: expert
  • MD5: c58949fab41e5bcbf0f0f4b26be306e5
  • Run description: Document organization and visualization based on Self Organized Maps initialized using actual CSIRO web site.

LiaIIcAuto

Results | Participants | Proceedings | Input | Summary

  • Run ID: LiaIIcAuto
  • Participant: lia-talne
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: f64627444ef29b876d848610a8558c05
  • Run description: Automatic run using SOM document similarity for query expansion

LiaIndriMan

Results | Participants | Proceedings | Input | Summary

  • Run ID: LiaIndriMan
  • Participant: lia-talne
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: manual
  • Task: document
  • MD5: 53686cd6f1d7dbe0075d353368b68a7e
  • Run description: Manual run using advanced features of indri language

LIAIndriSiac

Results | Participants | Proceedings | Input | Summary

  • Run ID: LIAIndriSiac
  • Participant: lia-talne
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 05c06b61358d2ec06edb3b852204d1c9
  • Run description: Automatic run using Indri and a tool used for QA (SIAC4QA).

pristask101

Results | Participants | Input | Summary

  • Run ID: pristask101
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/3/2008
  • Type: automatic
  • Task: document
  • MD5: 76a73bba72df969d9c31f7f88dd5a263
  • Run description: using the normal information retrevial method by adding the rights of the term from titles and the archor texts

pristask102

Results | Participants | Input | Summary

  • Run ID: pristask102
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/3/2008
  • Type: manual
  • Task: document
  • MD5: d4bedafbf14643576b6f7adf0798e0e6
  • Run description: using normal information retrieval method by the changed query which is choosen manully

pristask103

Results | Participants | Input | Summary

  • Run ID: pristask103
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/3/2008
  • Type: automatic
  • Task: document
  • MD5: 3627fb0ff28e4650ddd28d40005e1754
  • Run description: using the normal information retrival method by adding the rights of the terms from the titles only

pristask104

Results | Participants | Input | Summary

  • Run ID: pristask104
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/3/2008
  • Type: manual
  • Task: document
  • MD5: d23fa01af510d4d1ff1078cb537cc052
  • Run description: using the normal information retrieval method by the changed query,but only adding the right of the terms from title text

pristask201

Results | Participants | Input | Summary

  • Run ID: pristask201
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: expert
  • MD5: 207f82b448508134a1e3052cd9e28666
  • Run description: Using a window, which contains the information of an expert, as a document. Then the retrieval results contains the experts.

pristask202

Results | Participants | Input | Summary

  • Run ID: pristask202
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: automatic
  • Task: expert
  • MD5: bba85408c3aea30b1d1699f4d927dd4d
  • Run description: repeat the query 3 times. others is the first submitted files

pristask203

Results | Participants | Input | Summary

  • Run ID: pristask203
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: manual
  • Task: expert
  • MD5: 97985c947329248670e76a6253f6176a
  • Run description: review the query. others is the first submitted files

pristask204

Results | Participants | Input | Summary

  • Run ID: pristask204
  • Participant: BUPT_pris_
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: manual
  • Task: expert
  • MD5: 37efcd9cbf8cc58a490cb29c3acc6cee
  • Run description: review the query

Rkylv

Results | Participants | Proceedings | Input | Summary

  • Run ID: Rkylv
  • Participant: inria
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: 4a55e10fa0143ebc46ffe018fe7eb8bc

RmitDocQ

Results | Participants | Proceedings | Input | Summary

  • Run ID: RmitDocQ
  • Participant: rmit
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 208e398c791097c5f1c172462e02a516
  • Run description: RmitDocQ This this base run - Okapi ranking model with queries from query fields.

RmitDQCombLO

Results | Participants | Proceedings | Input | Summary

  • Run ID: RmitDQCombLO
  • Participant: rmit
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: a9f9d3819976e4679c19607d39777438
  • Run description: RmitDQComLO retrieved documents from RmitDocQ are re-ranked based on a linear combination of document weight and local out-links.

RmitDQExp

Results | Participants | Proceedings | Input | Summary

  • Run ID: RmitDQExp
  • Participant: rmit
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: e719758a1b6dc7d292743e258792c36b
  • Run description: RmitDQExp Pseudo relevance feedback by taking top 10 retrieved documents from rum RmitDocQ as "relevant" documents.

RmitDQRerank

Results | Participants | Proceedings | Input | Summary

  • Run ID: RmitDQRerank
  • Participant: rmit
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 18af5ea43ad96b24e4af4dc560b36a30
  • Run description: RmitDocQRerank top 100 retrieved documents from RmitDocQ run are re-ranked by using local out-links.

THUFaAS

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUFaAS
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 43299008b69ea04f6dac4004b655828e
  • Run description: Extract and integrate inlink anchor text for each document, and combine them into original documnets. Ranking with probablistic relevance model on the new collection.

THUFmfS

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUFmfS
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 0d0a34bca6498596da53a132c534a03b
  • Run description: Data clean for key resource selection. Ranking combination with probablistic relevance model on fulltext and cleaned data.

THUFS

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUFS
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 9294073ed0d752ebb466785299bd456a
  • Run description: baseline result. probablistic relevance model on fulltext.

THUFsimAncL

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUFsimAncL
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: c9cb568a6f801f613f4b13dcc0dcdb7a
  • Run description: Probabilistic relevance model on fulltext. Then similarity propagation with inlink docs. Finally re-ranking with anchor text similarity evidence.

THUPDDlchrS

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUPDDlchrS
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: 35c13f36656b474630b47769c5f4c7db
  • Run description: Combination of four runs Baseline result(PDD); Experts retrieved from the result of document search task, using link analysis to improve; PDD built with selected high quality pages; PDD built with experts' homepages. Topics about recruitment information are processed in particular.

THUPDDlcS

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUPDDlcS
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: bee218817b2379aa3cc654bd2497b8ba
  • Run description: Combination of three runs Baseline result(PDD); Experts retrieved from the result of document search task, using link analysis to improve; PDD built with selected high quality pages.

THUPDDSlL

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUPDDSlL
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: f9c59286d20a8e1d946f37f22373f3f2
  • Run description: Combination of two runs Baseline result(PDD); Experts retrieved from the result of document search task, using link analysis to improve. Narrative fields were used.

THUPDDSwp

Results | Participants | Proceedings | Input | Summary

  • Run ID: THUPDDSwp
  • Participant: THUIR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: f730762160ccfad82ed30c77f269c371
  • Run description: Baseline result document refinement to construct person description documents(PDD). experts search on PDD with word-pair-based probabilistic relevance ranking model.

TitBrf

Results | Participants | Proceedings | Input | Summary

  • Run ID: TitBrf
  • Participant: sebir
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: 08a71b279f1687dd51f596bb8692afa9
  • Run description: blind relevance feedback

TitDes

Results | Participants | Proceedings | Input | Summary

  • Run ID: TitDes
  • Participant: sebir
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: 6795bccde31eccec6495b376fde08335
  • Run description: title and description

TitExp

Results | Participants | Proceedings | Input | Summary

  • Run ID: TitExp
  • Participant: sebir
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: 3c2aa75ea55b11490d3e770036a3afe2
  • Run description: use wikipedia for query expansion

TitExpBrf57

Results | Participants | Proceedings | Input | Summary

  • Run ID: TitExpBrf57
  • Participant: sebir
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: 97c39dd9ddba29c558dca8bf61284fd6
  • Run description: use wikipedia for query expansion, query target collection with blind relevance feedback

TOmUW

Results | Participants | Proceedings | Input | Summary

  • Run ID: TOmUW
  • Participant: inria
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: c54d6d5e5bbbc21aae9dc621a9ca8efb

U2LwQ

Results | Participants | Proceedings | Input | Summary

  • Run ID: U2LwQ
  • Participant: st.petersburg
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: eada34985db83e25267b7ab18e6d47af

ualr08e01

Results | Participants | Input | Summary

  • Run ID: ualr08e01
  • Participant: UALR
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: d52b19e415faf5451d293b77f919ccfb
  • Run description: base run using combine, feedback, stemming, stop_words, and two-stage smoothing

ualr08e02

Results | Participants | Input | Summary

  • Run ID: ualr08e02
  • Participant: UALR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/3/2008
  • Type: automatic
  • Task: document
  • MD5: fd3aa44af514c25d17db119494258bc0
  • Run description: base run 2 using combine, feedback, stemming, stop_words, and two-stage smoothing. Main difference is change to mu.

ualr08e03

Results | Participants | Input | Summary

  • Run ID: ualr08e03
  • Participant: UALR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/3/2008
  • Type: automatic
  • Task: document
  • MD5: d3c2d74cf206fd8608f8c131b4d49144
  • Run description: run 3 used feedback, stemming, stop_words, two-stage smoothing, applied through #or(#band() #band()...) combinations, and powersets to provide different combinations of query words.

ualr08e04

Results | Participants | Input | Summary

  • Run ID: ualr08e04
  • Participant: UALR
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/3/2008
  • Type: manual
  • Task: document
  • MD5: c15d0dcd9cef84e42162ec5a6f19788f
  • Run description: used stemming, stop_words, two-stage smoothing, and utilized boolean queries generated by previous run through a perl script that submitted them to the index in fragments, sorted them hierarchically by indri relevance evaluation, and finally printed top the 1000 results.

ucl01

Results | Participants | Proceedings | Input | Summary

  • Run ID: ucl01
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: e6f6c4fa20aefe3356d2e87173530969
  • Run description: Language model Dirichlet smoothing, mu=2000.

ucl02

Results | Participants | Proceedings | Input | Summary

  • Run ID: ucl02
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: f653a1040e10bd97584657b73930f07c
  • Run description: Language model Dirichlet smoothing, mu=3000.

ucl03

Results | Participants | Proceedings | Input | Summary

  • Run ID: ucl03
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: 308d7fc5e0a769c29d05bb3bcf490e0d
  • Run description: Language model Dirichlet smoothing, mu=2500.

ucl04

Results | Participants | Proceedings | Input | Summary

  • Run ID: ucl04
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: document
  • MD5: 1b42785ccbd2b798f5ca9436bf7999be
  • Run description: BM25, K1=1.4, b=0.6, K3=8

UCLex01

Results | Participants | Proceedings | Input | Summary

  • Run ID: UCLex01
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: automatic
  • Task: expert
  • MD5: db55b0bd48b6a5a2287c5e52f77d65fc
  • Run description: Window size 450, anchor text, in-link, out-link

UCLex02

Results | Participants | Proceedings | Input | Summary

  • Run ID: UCLex02
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/13/2008
  • Type: automatic
  • Task: expert
  • MD5: 05f500cf1daa9c86174c7ce13d38453b
  • Run description: Window size 600, anchor text, in-link, out-link

UCLex03

Results | Participants | Proceedings | Input | Summary

  • Run ID: UCLex03
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/13/2008
  • Type: automatic
  • Task: expert
  • MD5: 9e4e9089f0e3739bd10ab0f53df34924
  • Run description: Multiple windows 40, 400, and 800, anchor text, in-link, out-link, PageRank, and document interal structures

UCLex04

Results | Participants | Proceedings | Input | Summary

  • Run ID: UCLex04
  • Participant: UCL
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/13/2008
  • Type: automatic
  • Task: expert
  • MD5: 66d0ffb927d34a61068a1711777770ce
  • Run description: Multiple windows 40, 200, 400, and 800, anchor text, in-link, out-link, PageRank, and document interal structures

uogTrEDbl

Results | Participants | Input | Summary

  • Run ID: uogTrEDbl
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 2ec96ac51db053538bcd3cc7a24f7684
  • Run description: fields and proximity

uogTrEDQE

Results | Participants | Input | Summary

  • Run ID: uogTrEDQE
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 5d351013d2b6c4c374fba4af382660e7
  • Run description: fields and query expansion

uogTrEDSE2

Results | Participants | Input | Summary

  • Run ID: uogTrEDSE2
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: e68e62acfff69d780d181e018684c058
  • Run description: selective query expansion and collection enrichment; Yahoo! Web Search was used as external resources

uogTrEDSelW

Results | Participants | Input | Summary

  • Run ID: uogTrEDSelW
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: a3d46732df5348de8ddf6adcc9d600b1
  • Run description: selective query expansion and collection enrichment; wikipedia was used as external resources

uogTrEXfeNP

Results | Participants | Input | Summary

  • Run ID: uogTrEXfeNP
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: 3357390a2cd65e0fdd60b350516dc6ca
  • Run description: voting model, query term proximity

uogTrEXfeNPC

Results | Participants | Input | Summary

  • Run ID: uogTrEXfeNPC
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: fa03f38973d03ffea62ca6c90fe1917f
  • Run description: voting model, candidate query term proximity

uogTrEXfePC

Results | Participants | Input | Summary

  • Run ID: uogTrEXfePC
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: 444f2a2b4983faacbc0489c34190e4b7
  • Run description: voting model, candidate query term proximity

uogTrEXmix

Results | Participants | Input | Summary

  • Run ID: uogTrEXmix
  • Participant: UoGtr
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: 2c595b66013cfb718d23bfaeb23dc76a
  • Run description: internal and external resources, voting model

utbase

Results | Participants | Input | Summary

  • Run ID: utbase
  • Participant: twentedb
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: expert
  • MD5: 243b611703270a8d486f1a84e1e7e97e
  • Run description: The submitted ranking is a combination of two rankings the ranking built using random walk based measure of expertise and the enterprise (CERC) data only, and the ranking using YahooAPI to find expertise evidence on the Web for each candidate from the baseline ranking.

utqtitle

Results | Participants | Input | Summary

  • Run ID: utqtitle
  • Participant: twentedb
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: expert
  • MD5: d68b9492d8f056a16f2686adeb00b348
  • Run description: The submitted ranking is a combination of two rankings the ranking built using random walk based measure of expertise and the enterprise (CERC) data only, and the ranking using YahooAPI to find expertise evidence on the Web for each candidate from the baseline ranking. In this run the relevance of the URL's title is taken into account.

utqurl

Results | Participants | Input | Summary

  • Run ID: utqurl
  • Participant: twentedb
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: expert
  • MD5: a476a29989ca8cdb185bd58985da15c9
  • Run description: The submitted ranking is a combination of two rankings the ranking built using random walk based measure of expertise and the enterprise (CERC) data only, and the ranking using YahooAPI to find expertise evidence on the Web for each candidate from the baseline ranking. In this run the relevance of the URL itself is taken into account.

utrecent

Results | Participants | Input | Summary

  • Run ID: utrecent
  • Participant: twentedb
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/5/2008
  • Type: automatic
  • Task: expert
  • MD5: b95a022633b63d265672cc2c1ed1cdfb
  • Run description: The submitted ranking is a combination of two rankings the ranking built using random walk based measure of expertise and the enterprise (CERC) data only, and the ranking using YahooAPI to find expertise evidence on the Web for each candidate from the baseline ranking. In this version of web based ranking only URLs with content dated after 2006 are used.

UvA08DSall

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08DSall
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 9848f945581a94ee516a852ed90339ad
  • Run description: UvA08DSexp + document priors based on URL character length

UvA08DSbfb

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08DSbfb
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 5536840c23b7a364aeb292ceff206935
  • Run description: Standard Language Modeling, ranking by query likelihood, plus blind relevance feedback

UvA08DSbl

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08DSbl
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: 4a8e92bb61a6aafb45fdd260bbfbf5a3
  • Run description: Standard Language Modeling, ranking by query likelihood

UvA08DSexp

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08DSexp
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/4/2008
  • Type: automatic
  • Task: document
  • MD5: f9d69fd31caf788407b24230777fa29e
  • Run description: UvA08DSbfb + query expansion using terms sampled from expert profiles

UvA08EScomb

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08EScomb
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: automatic
  • Task: expert
  • MD5: a68fa1c0bb15def27249921ab1f9e812
  • Run description: combination of Model 1B and Model 2, baseline query model

UvA08ESm1b

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08ESm1b
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: automatic
  • Task: expert
  • MD5: 768d7cb82307dc2c22914120ae0f46e3
  • Run description: Model 1B, semantic document-candidate associations, baseline query model

UvA08ESm2all

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08ESm2all
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: automatic
  • Task: expert
  • MD5: 629a263cdb59e5ddf6431386e18f7b1b
  • Run description: Model 2, query expansion using blind relevance feedback + expansion terms from expert profiles, document priors based on url length

UvA08ESweb

Results | Participants | Proceedings | Input | Summary

  • Run ID: UvA08ESweb
  • Participant: UAms_De_Rijke
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/12/2008
  • Type: automatic
  • Task: expert
  • MD5: c20c98c08cb3eb67faae582ffa1a3234
  • Run description: combination of Model 1B and Model 2, plus snippets from top 100 web search results, using YahooAPI.

WHU08BASE

Results | Participants | Proceedings | Input | Summary

  • Run ID: WHU08BASE
  • Participant: WHU
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: f08cf218c2da7b0c8afd0d0c71ded2ba
  • Run description: This is our baseline run. A combination of Bi-gram and uni-gram model are used for estimation p(q|d), while p(d|e) is simplified to 1. Besides, we use automatic ways of detecting possible phrases according to term statistical features in the corpus.

WHU08CAN

Results | Participants | Proceedings | Input | Summary

  • Run ID: WHU08CAN
  • Participant: WHU
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/13/2008
  • Type: automatic
  • Task: expert
  • MD5: 6c998cf0b9e803c27ae521bb2d7e3c56
  • Run description: In WHU08RFCAN, we use a refined method to estimate association between experts and documents. We consider how much each evidence can stands for a candidate. For most of the full name evidence and email evidence, they stand for the experts without any ambiguity. But for other ambiguous evidence, we estimate p(ca|evidence) to further work out the association. The WHU08CAN run does not use the associations stated above, and is submitted to make a comparison to WHU08RFCAN.

WHU08NOPHR

Results | Participants | Proceedings | Input | Summary

  • Run ID: WHU08NOPHR
  • Participant: WHU
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/13/2008
  • Type: automatic
  • Task: expert
  • MD5: f3d19419e0934cf5ce739df05868e760
  • Run description: A main feature of our methods this year is the automatic phrase detection in the query. This run is different from WHU08RFCAN in that it do not use the detection. We submit this run to make a comparison, though effect of phrase detection is already proved by us in experiments using TREC 07 queries.

WHU08RFCAN

Results | Participants | Proceedings | Input | Summary

  • Run ID: WHU08RFCAN
  • Participant: WHU
  • Track: Enterprise
  • Year: 2008
  • Submission: 8/11/2008
  • Type: automatic
  • Task: expert
  • MD5: fccc5886b1ce8de3606daa4e68d9901e
  • Run description: Compared with the baseline run, this run do not simplified candidate-document associations. The ambiguity of each evidence is considered in this estimation and received better results in TREC 07's collections.

xLQOW

Results | Participants | Proceedings | Input | Summary

  • Run ID: xLQOW
  • Participant: st.petersburg
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: 68ba9a10dd4016285e0875587aef098a

ycbLS

Results | Participants | Proceedings | Input | Summary

  • Run ID: ycbLS
  • Participant: inria
  • Track: Enterprise
  • Year: 2008
  • Submission: 7/30/2008
  • Type: automatic
  • Task: document
  • MD5: 75a21b6ef8d0c4c58840be4ed4916b47