Runs - Web 2011¶

2011SiftR1¶

Run ID: 2011SiftR1
Participant: UCDSIFT
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 83c2bd8bc04cca865fa2c05e4130361e
Run description: Fusion of Indri runs using SlideFuse & 2 years training data

2011SiftR2¶

Run ID: 2011SiftR2
Participant: UCDSIFT
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: b94330e5444b6be9220b419c705377a3
Run description: Fusion of Indri runs using SlideFuse & 1 year of training data

2011SiftR3¶

Run ID: 2011SiftR3
Participant: UCDSIFT
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: d442be9f1a14deddc4fc350b4a49ffb5
Run description: Fusion of Terrier runs using SlideFuse & 2 years training data

bpacad1¶

Run ID: bpacad1
Participant: budapest_acad
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 302c01e58db040914a16a8585309c01a
Run description: Okapi BM25 ranking methode

bpacadspam1¶

Run ID: bpacadspam1
Participant: budapest_acad
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 7b0af703d121ce10b3be9e3b33b376d1
Run description: Okapi BM25 ranking + webspam filtering(svm)

bpacadspam2¶

Run ID: bpacadspam2
Participant: budapest_acad
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 22b7e3c491c8bb9a669883c46b2730c4
Run description: Okapi BM25 ranking + spam filtering (svm, threshold:0.3)

CWIcIA2t5b1¶

Run ID: CWIcIA2t5b1
Participant: CWI
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: f1128bf886028fd840b5b8d9c3405e0b
Run description: same as run CWIcIAt5b1, but using a different weighting scheme to weight the importance of a related query in a cluster.

CWIcIAt5b1¶

Run ID: CWIcIAt5b1
Participant: CWI
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: 69ec8438be296f2f21fc7bec2d06d4ba
Run description: -using query log to collect related queries for a given query - cluster the related queries, each cluster is an aspect of the original query - use IA-select to diversify

CWIIAt5b5¶

Run ID: CWIIAt5b5
Participant: CWI
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: f2501f3a57c8f200c35c27bf2d56ee7c
Run description: -using query log to collect related queries for a given query - using top10 related queries as aspects of the original query - use IA-select to diversify

DFalah11¶

Run ID: DFalah11
Participant: uottawa
Track: Web
Year: 2011
Submission: 7/25/2011
Task: adhoc
MD5: eda047245b0ffa0d2e88482bc51a79df
Run description: Our system exploits the collective knowledge from Wikipedia articles and from Alexa IC. We extracted relevant documents and related topics from Wikipedia, relevant documents from external links, and relevant topics similar to internal topics. Also, we used some data from Alexa IC to filter out spam and irrelevant documents (with low reputation and high traffic).

ICTNET11ADR2¶

Run ID: ICTNET11ADR2
Participant: ICTNET
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: acfefa0eb342f7860f59ebc119473806
Run description: In this run, we first pre-process the query with query expansion and entity extraction. Then use BM25 sliding window model on content field.

ICTNET11ADR3¶

Run ID: ICTNET11ADR3
Participant: ICTNET
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 098d4b796374eeb0bcc9705b5445e101
Run description: In this run, we diversify ICTNET11ADR2.

ICTNET11ADR4¶

Run ID: ICTNET11ADR4
Participant: ICTNET
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 7a427e3221094d4b1db74810076a2b0d
Run description: In this run, we use an combination of Title field and external anchor data provided by http://lemurproject.org/clueweb09/anchortext-querylog/ For the Title field, we use BM25 score with operator OR. For the external anchor data, we use BM25 score and sorting by count DESC. Also, we consider dmoz.org top 1 and Bing.com top 10 as authoritative site, and boost the result.

ICTNET11DVR1¶

Run ID: ICTNET11DVR1
Participant: ICTNET
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: 23c784f0f2cabf5c39984b4d0acaa8e6
Run description: This run used google query expansion.

ICTNET11DVR2¶

Run ID: ICTNET11DVR2
Participant: ICTNET
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: a7976429df9438ff2835811a9857ae66
Run description: This run used google query expansion.

ICTNET11DVR3¶

Run ID: ICTNET11DVR3
Participant: ICTNET
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: 224a170c26a51a3cb84231bac274e67b
Run description: This run used google and bing query expansion.

liaQEWikiA¶

Run ID: liaQEWikiA
Participant: LIA
Track: Web
Year: 2011
Submission: 7/24/2011
Task: adhoc
MD5: d537afb165acc9ae3ec19210a7ce3e3a
Run description: A run that uses Wikipedia pages obtained with its search APIs. Terms are extracted from the first Wikipedia page, and then added to the original query in a pseudo-relevance feedback fashion. These terms are appropriately weighted and incorporated in a language model for retrieval.

liaQEWikiAnA¶

Run ID: liaQEWikiAnA
Participant: LIA
Track: Web
Year: 2011
Submission: 7/24/2011
Task: diversity
MD5: 3d30833f049088be48218ac994b5b8f6
Run description: A run that uses Wikipedia pages obtained with its search APIs. Terms are extracted from the first Wikipedia page, and then added to the original query in a pseudo-relevance feedback fashion. A thematic graph is built using the anchor links of the Wikipedia page, and other terms are extracted from linked pages. All these terms are appropriately weighted and incorporated in a language model for retrieval.

liaQEWikiGoA¶

Run ID: liaQEWikiGoA
Participant: LIA
Track: Web
Year: 2011
Submission: 7/24/2011
Task: adhoc
MD5: de11d2db92e38b0963d41c8fc77e460b
Run description: A run that uses Wikipedia and Google pages obtained with their search APIs. Terms are extracted from the first Wikipedia and Google pages, and then added to the original query in a pseudo-relevance feedback fashion. These terms are appropriately weighted and incorporated in a language model for retrieval.

liaQEWikiGoo¶

Run ID: liaQEWikiGoo
Participant: LIA
Track: Web
Year: 2011
Submission: 7/24/2011
Task: adhoc
MD5: 458d2525252d11c6b9266b69ba8019d0
Run description: A run that uses Wikipedia and Google pages obtained with their search APIs. Terms are extracted from the first Wikipedia and Google pages, and then added to the original query in a pseudo-relevance feedback fashion. These terms are appropriately weighted and incorporated in a language model for retrieval.

msrsv2011a1¶

Run ID: msrsv2011a1
Participant: msrsv
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 556e1be7021fd89bb23850a8ea25b486
Run description: Five features (MAC,BM25F,SALSA,IDID,Spam) x two anchor-based query rewriting techniques, weighted linear combination of evidence. Additional resources: DNS resolutions

msrsv2011a2¶

Run ID: msrsv2011a2
Participant: msrsv
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: e4cbc58e730c1afb0b6c8ec62e6250de
Run description: Five features (MAC,BM25F,SALSA,IDID,Spam) x two anchor-based query rewriting techniques, weighted linear combination of evidence. Additional resources: DNS resolutions

msrsv2011a3¶

Run ID: msrsv2011a3
Participant: msrsv
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: da6ec85304c4639ed671bfc3e1b9be2c
Run description: Five features (MAC,BM25F,SALSA,IDID,Spam) x two anchor-based query rewriting techniques, weighted linear combination of evidence. Additional resources: DNS resolutions

msrsv2011d1¶

Run ID: msrsv2011d1
Participant: msrsv
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: a7b4d0d2d59360025f210e35be0a1798
Run description: Five features (MAC,BM25F,SALSA,IDID,Spam) x two anchor-based query rewriting techniques, weighted linear combination of evidence. Additional resources: DNS resolutions

msrsv2011d2¶

Run ID: msrsv2011d2
Participant: msrsv
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: dd971c5afc01e2c0b97d9819699fe296
Run description: Five features (MAC,BM25F,SALSA,IDID,Spam) x two anchor-based query rewriting techniques, weighted linear combination of evidence. Additional resources: DNS resolutions

msrsv2011d3¶

Run ID: msrsv2011d3
Participant: msrsv
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: e1229d82853da65687c19f83f4438dfe
Run description: Five features (MAC,BM25F,SALSA,IDID,Spam) x two anchor-based query rewriting techniques, weighted linear combination of evidence. Additional resources: DNS resolutions

Otago2011cn¶

Run ID: Otago2011cn
Participant: Otago
Track: Web
Year: 2011
Submission: 7/20/2011
Task: adhoc
MD5: dfc128c1d2699d00285ffab4d69bf390
Run description: Category B. Central Nothing. Distributed search engine default parameters

Otago2011cnw¶

Run ID: Otago2011cnw
Participant: Otago
Track: Web
Year: 2011
Submission: 7/20/2011
Task: adhoc
MD5: 461d07eafa4cd51519b2b02533c73b93
Run description: Categoey B Wikipedia only run. Central nothing. Default parameters to serch engine.

qutir11a¶

Run ID: qutir11a
Participant: QUT_IR
Track: Web
Year: 2011
Submission: 7/25/2011
Task: adhoc
MD5: 6d311cc3c49a5557a8a0ecce54dcbd40
Run description: This run uses signatures and pseudorelevance feedback to rank documents in the ClueWeb09 Category B collection. The additional resource used was the Waterloo 'fusion' Spam Ranking file, although its influence on the output was very low.

qutir11b¶

Run ID: qutir11b
Participant: QUT_IR
Track: Web
Year: 2011
Submission: 7/26/2011
Task: adhoc
MD5: d01d380353d4024de1e640ef535082b2
Run description: Same as the qutir11a run, except filtering on a smaller pool because this seems to produce better results on the 2010 topics + qrels. Again it uses the 'fusion' spam table (and actually gives it a slightly higher weight this time.)

srchvrs11b¶

Run ID: srchvrs11b
Participant: srchvrs
Track: Web
Year: 2011
Submission: 7/19/2011
Task: adhoc
MD5: 0c266ccaa978d6897e558361bed24a23
Run description: Waterloo University spamrank, CMU PageRank.

srchvrs11o¶

Run ID: srchvrs11o
Participant: srchvrs
Track: Web
Year: 2011
Submission: 7/19/2011
Task: adhoc
MD5: 04b9b2a0072535f592228318b3b60bda
Run description: Waterloo University spamrank.

UAmsAnc05LS¶

Run ID: UAmsAnc05LS
Participant: UAmsterdam
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: 689c2e62056ce1de18ebc3e8cb4c2524
Run description: UAmsAnc05LS is a run on an anchor-text index of ClueWeb09 B, using a language model with Jelinek-Mercer smoothing and lambda=0.95. Two document priors are used. A length prior based on the character length of the anchor-text representation, and a spam prior based on the Spam rankings provided by Waterloo. The spam prior uses the spam percentile as a prior score.

UAmsM705FLS¶

Run ID: UAmsM705FLS
Participant: UAmsterdam
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: d00e0fd08db7f2eff25ca5709bd18b78
Run description: UAmsM705FLS is a run based on a fusion of 70% full-text (with relevance feedback) and 30% anchor-text index runs on ClueWeb09 B, using a language model with JM smoothing and lambda=0.05. A length prior is used on the anchor-text and full-text indexes based on character length of the document representation. After fusion, a spam prior is used based on the Spam rankings provided by Waterloo. The spam prior uses the spam percentile as a prior score.

UAmsM705tFLS¶

Run ID: UAmsM705tFLS
Participant: UAmsterdam
Track: Web
Year: 2011
Submission: 7/28/2011
Task: diversity
MD5: 73e99b42727a5bbd4ebaab995d2e037d
Run description: UAmsM705tiLS is a run based on a fusion of 70% full-text (with relevance feedback) and 30% anchor-text index runs on ClueWeb09 B, using a language model with Jelinek-Mercer smoothing and lambda=0.95. On the full-text index, query terms occurring in the document title receive extra weight. Two document priors are used. A length prior based on the character length of the anchor-text representation, and a spam prior based on the Spam rankings provided by Waterloo. The spam prior uses the spam percentile as a prior score.

UAmsM705tiLS¶

Run ID: UAmsM705tiLS
Participant: UAmsterdam
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 38b4cfe6ac92c96cca81816dadd6156e
Run description: UAmsM705tiLS is a run based on a fusion of 70% full-text and 30% anchor-text index runs on ClueWeb09 B, using a language model with Jelinek-Mercer smoothing and lambda=0.95. On the full-text index, query terms occurring in the document title receive extra weight. Two document priors are used. A length prior based on the character length of the anchor-text representation, and a spam prior based on the Spam rankings provided by Waterloo. The spam prior uses the spam percentile as a prior score.

UAmsM7DirExS¶

Run ID: UAmsM7DirExS
Participant: UAmsterdam
Track: Web
Year: 2011
Submission: 7/27/2011
Task: diversity
MD5: 4da4467b16cbe459d77633b56198cafa
Run description: UAmsM7DirExS is a run based on a fusion of 70% full-text and 30% anchor-text index runs on ClueWeb09 B, using a language model with Dirichlet smoothing and mu=2500. On the full-text index, query terms occurring in the document title receive extra weight. A prior is used on the anchor-text and full-text indexes based on character length of the document representationl. After fusion, an extreme a spam prior is used based on the Spam rankings provided by Waterloo. The spam prior uses the square of the spam percentile as a prior score.

UAmsT05FLS¶

Run ID: UAmsT05FLS
Participant: UAmsterdam
Track: Web
Year: 2011
Submission: 7/27/2011
Task: adhoc
MD5: 217d907352996a055b069469ff6b2552
Run description: UAmsM705FLS is a run based on a full-text with relevance feedback on ClueWeb09 B, using a language model with JM smoothing and lambda=0.05. A length prior is used based on character length of the document representation. A spam prior is used based on the Spam rankings provided by Waterloo, using the spam percentile as a prior score.

UDCombine1¶

Run ID: UDCombine1
Participant: Udel_Fang
Track: Web
Year: 2011
Submission: 7/21/2011
Task: diversity
MD5: a2bb43cab3f049ebcef81e8ca68addff
Run description: We extract subtopics using integration information of PLSA clusters and ODP. The diversification function is SQR.

UDCombine2¶

Run ID: UDCombine2
Participant: Udel_Fang
Track: Web
Year: 2011
Submission: 7/21/2011
Task: diversity
MD5: 5984701d5fd152b76aa4449d07a9e437
Run description: We extract subtopics using integration information of PLSA clusters and patterns. The diversification function is SQR.

UDPattern¶

Run ID: UDPattern
Participant: Udel_Fang
Track: Web
Year: 2011
Submission: 7/21/2011
Task: diversity
MD5: 7a8734db024f459fb451c81533bc02d3
Run description: We extract subtopics using the pattern-based method and use SQR to diversify results based on these subtopics.

umassGFuse¶

Run ID: umassGFuse
Participant: CIIR
Track: Web
Year: 2011
Submission: 7/21/2011
Task: adhoc
MD5: 1b80c2dd4dd4989ba0efd27203e7b092
Run description: Sequential dependence and query expansion. Using Galago search engine.

umassGQdist¶

Run ID: umassGQdist
Participant: CIIR
Track: Web
Year: 2011
Submission: 7/21/2011
Task: adhoc
MD5: 5dd5b11f479129e3183547077e5855d3
Run description: Query Distributions. Using Galago search engine.

umassIFuse¶

Run ID: umassIFuse
Participant: CIIR
Track: Web
Year: 2011
Submission: 7/21/2011
Task: adhoc
MD5: 19ee780a5c2508ab07e9f2250c15d25b
Run description: Sequential dependence and query expansion. Using Indri search engine.

uogTrA45Nm¶

Run ID: uogTrA45Nm
Participant: uogTr
Track: Web
Year: 2011
Submission: 7/20/2011
Task: adhoc
MD5: 8a8f8ad424e2acb34d35bb26e8a51b3a
Run description: Learned model including various Divergence from Randomness and other features.

uogTrA45Nmx2¶

Run ID: uogTrA45Nmx2
Participant: uogTr
Track: Web
Year: 2011
Submission: 7/21/2011
Task: diversity
MD5: c0fc80cc3326c225e2abe6e9453153ca
Run description: Learned diversification model

uogTrA45Vm¶

Run ID: uogTrA45Vm
Participant: uogTr
Track: Web
Year: 2011
Submission: 7/20/2011
Task: adhoc
MD5: 44db80561f3e5c513a7687a9d8199f54
Run description: Learned model including various Divergence from Randomness and other features.

uogTrA45Vmx¶

Run ID: uogTrA45Vmx
Participant: uogTr
Track: Web
Year: 2011
Submission: 7/21/2011
Task: diversity
MD5: b9d3d8e9f88198187a7596f8ff39a855
Run description: Learned diversification model

uogTrB47Vm¶

Run ID: uogTrB47Vm
Participant: uogTr
Track: Web
Year: 2011
Submission: 7/20/2011
Task: adhoc
MD5: c6dda035b91a38c83d941fcd4b006cb5
Run description: Learned model including various Divergence from Randomness and other features.

uogTrB47Vmx¶

Run ID: uogTrB47Vmx
Participant: uogTr
Track: Web
Year: 2011
Submission: 7/21/2011
Task: diversity
MD5: 4e2c3edd303f839e5597f82b62fea7ff
Run description: Learned diversification model

UWatMDSdm¶

Run ID: UWatMDSdm
Participant: UWaterlooMDS
Track: Web
Year: 2011
Submission: 7/23/2011
Task: adhoc
MD5: 035ef1fe2089d7d99802b98c3b0a258a
Run description: Dependence models. Waterloo spam filtered. Twente anchortext.

UWatMDSdmsr¶

Run ID: UWatMDSdmsr
Participant: UWaterlooMDS
Track: Web
Year: 2011
Submission: 7/23/2011
Task: diversity
MD5: e1c5d3092a7cf91c67d46252955cbf40
Run description: Dependence models with stopwords removed. Waterloo spam filtered. Twente anchortext.

UWatMDSql¶

Run ID: UWatMDSql
Participant: UWaterlooMDS
Track: Web
Year: 2011
Submission: 7/23/2011
Task: adhoc
MD5: 7939b2ed7ff7b85824e3360f83b5d1d4
Run description: Query likelihood. Waterloo spam filtered. Twente anchortext.

UWatMDSqlsr¶

Run ID: UWatMDSqlsr
Participant: UWaterlooMDS
Track: Web
Year: 2011
Submission: 7/23/2011
Task: diversity
MD5: 74c215aeea79104b9af589fcfe10690f
Run description: Query likelihood with stopwords removed. Waterloo spam filtered. Twente anchortext.

UWatMDSqlt¶

Run ID: UWatMDSqlt
Participant: UWaterlooMDS
Track: Web
Year: 2011
Submission: 7/23/2011
Task: adhoc
MD5: a48eed237eaf65239e99266f7c6043dc
Run description: Query likelihood with titles weighted. Waterloo spam filtered. Twente anchortext.

UWatMDSqltsr¶

Run ID: UWatMDSqltsr
Participant: UWaterlooMDS
Track: Web
Year: 2011
Submission: 7/23/2011
Task: diversity
MD5: 6d022765ead376b3d5021846afac4294
Run description: Query likelihood with titles weighted and with stopwords removed. Waterloo spam filtered. Twente anchortext.

uwBA¶

Run ID: uwBA
Participant: waterloo
Track: Web
Year: 2011
Submission: 7/26/2011
Task: diversity
MD5: bcff8f9f7fa4120af979d47f0e3a26d6
Run description: Retrieval was done on a spam-filtered index of ClueWeb Collection A.

uwBAadhoc¶

Run ID: uwBAadhoc
Participant: waterloo
Track: Web
Year: 2011
Submission: 7/26/2011
Task: adhoc
MD5: 38bd63460d81f1497b7a5a93a8177d29
Run description: Retrieval was done on a spam-filtered index of ClueWeb Collection A.

uwBBadhoc¶

Run ID: uwBBadhoc
Participant: waterloo
Track: Web
Year: 2011
Submission: 7/26/2011
Task: adhoc
MD5: 92cc3ce2f1edb44634fe1fab49e7fa37
Run description: Retrieval was done on a spam-filtered index of ClueWeb Collection B.

uwExpAa¶

Run ID: uwExpAa
Participant: waterloo
Track: Web
Year: 2011
Submission: 7/26/2011
Task: diversity
MD5: b901c0d8be7c0cf29f45b23b0d67b41d
Run description: Anchor text was used to discover diverse intents of the query. Terms that co-occur more frequently with the original query were clustered using their co-occurrence information in the anchor text collection. The original query was expanded with terms selected from the clusters. Results from each expanded queries were selected on a round-robin manner. Retrieval was done on a spam-filtered Clueweb category A collection.

uwExpAadhoc¶

Run ID: uwExpAadhoc
Participant: waterloo
Track: Web
Year: 2011
Submission: 7/26/2011
Task: adhoc
MD5: 4197c3216b74654e9f8a331374e0d4cf
Run description: Anchor text was used to discover diverse intents of the query. Terms that co-occur more frequently with the original query were clustered using their co-occurrence information in the anchor text collection. The original query was expanded with terms selected from the clusters. Results from each expanded queries were selected on a round-robin manner. Retrieval was done on a spam-filtered Clueweb category A collection.

uwExpB¶

Run ID: uwExpB
Participant: waterloo
Track: Web
Year: 2011
Submission: 7/26/2011
Task: diversity
MD5: 8881cedf8d329ad338ac1b42a9214dec
Run description: Anchor text was used to discover diverse intents of the query. Terms that co-occur more frequently with the original query were clustered using their co-occurrence information in the anchor text collection. The original query was expanded with terms selected from the clusters. Results from each expanded queries were selected on a round-robin manner.