Runs - Blog 2007¶

ADABoostM1¶

Results | Participants | Proceedings | Input | Summary

Run ID: ADABoostM1
Participant: uberlin.neubauer
Track: Blog
Year: 2007
Submission: 8/27/2007
Type: automatic
Task: feed
Run description: ADABoost.M1, Wordnet, Wikipedia, Dmoz, Yahoo, Ngram

BaseSmp¶

Run ID: BaseSmp
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: An automatic run, using stactics method.

CMUentry¶

Results | Participants | Proceedings | Input | Summary

Run ID: CMUentry
Participant: cmu.dir.callan
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Small document retrieval with Relevance Model PRF

CMUentryW¶

Results | Participants | Proceedings | Input | Summary

Run ID: CMUentryW
Participant: cmu.dir.callan
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Small document retrieval with Relevance Model PRF and Wikipedia link-based PRF

CMUfeed¶

Results | Participants | Proceedings | Input | Summary

Run ID: CMUfeed
Participant: cmu.dir.callan
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Large document retrieval with Relevance Model PRF

CMUfeedW¶

Results | Participants | Proceedings | Input | Summary

Run ID: CMUfeedW
Participant: cmu.dir.callan
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Large document retrieval with Relevance Model PRF and Wikipedia link-based expansion

DrapCom¶

Run ID: DrapCom
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: An automatic run, using our own push-drag method.

DrapComSub¶

Results | Participants | Proceedings | Input | Summary

Run ID: DrapComSub
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

DrapOpi¶

Run ID: DrapOpi
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: An automatic run, using our own lexicon based push-drag method.

DrapOpiSub¶

Results | Participants | Proceedings | Input | Summary

Run ID: DrapOpiSub
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

DrapStm¶

Run ID: DrapStm
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: An automatic run, stemming each doucument and using our own lexicon based push-drag method.

DrapStmSub¶

Results | Participants | Proceedings | Input | Summary

Run ID: DrapStmSub
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

DUTDRun1¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTDRun1
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/16/2007
Type: automatic
Task: feed
Run description: This is an automatic run which use title only.

DUTDRun2¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTDRun2
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/16/2007
Type: automatic
Task: feed
Run description: This is an automatic run which use title and description fields.

DUTDRun3¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTDRun3
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/16/2007
Type: automatic
Task: feed
Run description: This is an automatic run which use title,description and narrative fields.

DUTDRun4¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTDRun4
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/16/2007
Type: manual
Task: feed
Run description: This is an manual run which use title,description and narrative fields.

DUTRun1¶

Run ID: DUTRun1
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: This is a Topic-Relevance run, and we set all the documentlabel in polarity file to 0.

DUTRun1P¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTRun1P
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

DUTRun2¶

Run ID: DUTRun2
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run is completed only with title field and with opinion words added into the query.

DUTRun2P¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTRun2P
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

DUTRun3¶

Run ID: DUTRun3
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run is completed with title and description fields and with opinion words added into the query.

DUTRun3P¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTRun3P
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

DUTRun4¶

Run ID: DUTRun4
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run is completed with title and description fields and without opinion words added into the query.

DUTRun4P¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTRun4P
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

DUTRun5¶

Run ID: DUTRun5
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run is completed with title and description and narrative fields and without opinion words added into the query.

DUTRun5P¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTRun5P
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

DUTRun6¶

Run ID: DUTRun6
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: manual
Task: opinion
Run description: This run is completed manually with title and description and narrative fields and with opinion words added into the query.

DUTRun6P¶

Results | Participants | Proceedings | Input | Summary

Run ID: DUTRun6P
Participant: dalianu.yang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

EAGLE1¶

Run ID: EAGLE1
Participant: zhejiangu.qiu
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: The topic-relevance run (all opinion-finding features turned off)

EAGLE2¶

Run ID: EAGLE2
Participant: zhejiangu.qiu
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: Dictionary-based approach

EAGLE2P¶

Results | Participants | Input | Summary

Run ID: EAGLE2P
Participant: zhejiangu.qiu
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

EAGLE3¶

Run ID: EAGLE3
Participant: zhejiangu.qiu
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: Using Naive Bayes for classification

EAGLE3P¶

Results | Participants | Input | Summary

Run ID: EAGLE3P
Participant: zhejiangu.qiu
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

EAGLE4¶

Run ID: EAGLE4
Participant: zhejiangu.qiu
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: Using support vector machine for classification

EAGLE4P¶

Results | Participants | Input | Summary

Run ID: EAGLE4P
Participant: zhejiangu.qiu
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

FDUNoOpTisd¶

Run ID: FDUNoOpTisd
Participant: fudanu.wu
Track: Blog
Year: 2007
Submission: 8/3/2007
Type: automatic
Task: opinion
Run description: Query expansion is used in this run. It is a topic-relevance run with no opinion-finding module.

FDUNoOpTSem¶

Run ID: FDUNoOpTSem
Participant: fudanu.wu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Based on the run FDUNoOpTisd, a sentiment process is applied in this run.

FDUNOpRSVMT¶

Run ID: FDUNOpRSVMT
Participant: fudanu.wu
Track: Blog
Year: 2007
Submission: 8/3/2007
Type: automatic
Task: opinion
Run description: Query expansion is used in this run. The extended terms for a query are ordered by a SVM classifier. It is a topic-relevance run with no opinion-finding module.

FDUTisdOpSVM¶

Run ID: FDUTisdOpSVM
Participant: fudanu.wu
Track: Blog
Year: 2007
Submission: 8/3/2007
Type: automatic
Task: opinion
Run description: Query expansion is used in this run. A SVM classifier trained on the opinion relevant pool is used to rerank the extended terms. Opinion-finding features are used in this run.

FDUTNRSVMSem¶

Run ID: FDUTNRSVMSem
Participant: fudanu.wu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Based on the run FDUNOpRSVMT, a sentiment process is used in this run.

FDUTOSVMSem¶

Run ID: FDUTOSVMSem
Participant: fudanu.wu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Based on the run FDUTisdOpSVM, a sentiment process is used in this run.

FIUbPL2¶

Run ID: FIUbPL2
Participant: fondazione-ug-bordoni-netlab-team
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This is the baseline PL2 with all opinion-finding features turned off.

FIUDDPH¶

Run ID: FIUDDPH
Participant: fondazione-ug-bordoni-netlab-team
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: Rank computed using DPH with query expansion and title+description fields.

FIUdPL2¶

Run ID: FIUdPL2
Participant: fondazione-ug-bordoni-netlab-team
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: Rank rearranged by a sentimental score, computed by PL2 with respect to sentimental query.

FIUlDPH¶

Run ID: FIUlDPH
Participant: fondazione-ug-bordoni-netlab-team
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: As FIUlPL2 but DPH is used instead of PL2.

FIUlP22¶

Run ID: FIUlP22
Participant: fondazione-ug-bordoni-netlab-team
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: As FIUlPL2 but sentiment term scores are normalized using PL2 instead of DPH.

FIUlPL2¶

Run ID: FIUlPL2
Participant: fondazione-ug-bordoni-netlab-team
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: As FIUlPL2, but weighting the sentimental score by learning.

KobePrMIR01¶

Run ID: KobePrMIR01
Participant: kobeu.eguchi
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: We will let you know later

KobePrMIR02¶

Run ID: KobePrMIR02
Participant: kobeu.eguchi
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: We will let you know later

KobePrMIR03¶

Run ID: KobePrMIR03
Participant: kobeu.eguchi
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: We will let you know later

KobePrMIR04¶

Run ID: KobePrMIR04
Participant: kobeu.eguchi
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: We will let you know later

KobePrMIR05¶

Run ID: KobePrMIR05
Participant: kobeu.eguchi
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: (We will let you know later.)

KobePrMIR06¶

Run ID: KobePrMIR06
Participant: kobeu.eguchi
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: (We will let you know later.)

Ku¶

Run ID: Ku
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: search by topic title only

kud¶

Results | Participants | Proceedings | Input | Summary

Run ID: kud
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: feed search + permalink search

kudn¶

Results | Participants | Proceedings | Input | Summary

Run ID: kudn
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: feed search + permalink search + non-english blog exclusion

kuds¶

Results | Participants | Proceedings | Input | Summary

Run ID: kuds
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: feed search + permalink search + splog exclusion

kudsn¶

Results | Participants | Proceedings | Input | Summary

Run ID: kudsn
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: feed search + permalink search + non-english blog exclusion + splog exclusion

KuKnn¶

Run ID: KuKnn
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: search by topic title only. use knn to classify opinions and non-opinions.

KuS¶

Run ID: KuS
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: search by topic title only. include splog detection by svm.

KuSvm¶

Run ID: KuSvm
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: search by topic title only. use svm to classify opinons and non-opinions.

KuSvmS¶

Run ID: KuSvmS
Participant: kobeu.seki
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: search by topic title only. use svm to classify opinons and non-opinions. include splog detection by svm.

NLPRPST¶

Run ID: NLPRPST
Participant: cas-liu-nlpr-iacas
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: We use web information extraction method to clear web page and use title and as query. But sentence selecting is different to NLPRPT

NLPRPT¶

Run ID: NLPRPT
Participant: cas-liu-nlpr-iacas
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: We use web information extraction method to clear web page and use title field as query.

NLPRPTD1¶

Run ID: NLPRPTD1
Participant: cas-liu-nlpr-iacas
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: We use web information extraction method to clear web page and use title and description field as query

NLPRPTD2¶

Run ID: NLPRPTD2
Participant: cas-liu-nlpr-iacas
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: We use web information extraction method to clear web page and use title and description as query. It's parameter is different to NLPRPTD1

NLPRPTONLY¶

Run ID: NLPRPTONLY
Participant: cas-liu-nlpr-iacas
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: We use web information extraction method to clear web page and use title field as query. This result is topic-relevance run with all opinion-finding features turned off.

NLPRTD¶

Run ID: NLPRTD
Participant: cas-liu-nlpr-iacas
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: We don't use web information extraction method to clear web page and use title and description field as query

NOOPWHU1¶

Run ID: NOOPWHU1
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run is for opinion task with all opinion-finding feature turning off. It is based on OKAPI system.

NTUAuto¶

Run ID: NTUAuto
Participant: ntu.chen
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Topic-relevance-only run.

NTUAutoOp¶

Run ID: NTUAutoOp
Participant: ntu.chen
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Automated opinion retrieval run.

NTUAutoOpP¶

Results | Participants | Proceedings | Input | Summary

Run ID: NTUAutoOpP
Participant: ntu.chen
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

NTUManual¶

Run ID: NTUManual
Participant: ntu.chen
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: manual
Task: opinion
Run description: Topic-relevance-only run. Queries are manually extracted from title fields.

NTUManualOp¶

Run ID: NTUManualOp
Participant: ntu.chen
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: manual
Task: opinion
Run description: Queries are manually extracted from title fields. Document opinions are identified using SVM.

NTUManualOpP¶

Results | Participants | Proceedings | Input | Summary

Run ID: NTUManualOpP
Participant: ntu.chen
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

oqlnr2opt¶

Run ID: oqlnr2opt
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: okapi weight, topic & opinion reranking (optimized)

oqlnr2optP¶

Results | Participants | Proceedings | Input | Summary

Run ID: oqlnr2optP
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

oqlr2f2¶

Run ID: oqlr2f2
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: okapi weights, topic & opinion reranking, MAP fusion

oqlr2f2opt¶

Run ID: oqlr2f2opt
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: okapi weights, topic & opinion reranking (optimized), MAP fusion

oqlr2f2optP¶

Results | Participants | Proceedings | Input | Summary

Run ID: oqlr2f2optP
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

oqlr2f2P¶

Results | Participants | Proceedings | Input | Summary

Run ID: oqlr2f2P
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

oqlr2fopt¶

Run ID: oqlr2fopt
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: okapi weights, topic & opinion reranking (optimized), basic fusion

oqlr2foptP¶

Results | Participants | Proceedings | Input | Summary

Run ID: oqlr2foptP
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

oqsnr1Base¶

Run ID: oqsnr1Base
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: okapi weight, topic reranking (opinion features turned off)

oqsnr2opt¶

Run ID: oqsnr2opt
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: okapi weight, topic & opinion reranking (optimized)

oqsnr2optP¶

Results | Participants | Proceedings | Input | Summary

Run ID: oqsnr2optP
Participant: indianau.yang
Track: Blog
Year: 2007
Submission: 8/5/2007
Task: polarity

OTPSWHU101¶

Results | Participants | Proceedings | Input | Summary

Run ID: OTPSWHU101
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

OTPSWHU102¶

Results | Participants | Proceedings | Input | Summary

Run ID: OTPSWHU102
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

OTWHU101¶

Run ID: OTWHU101
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run is ranked based on the multiple of doc weight for Topics and Opinionated weight of each doc. It is based on Okapi system with stemmed Index.

OTWHU102¶

Run ID: OTWHU102
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run is ranked based on the doc weights for each Topic, posprocessed according to a certain threshold value of opinionated nature. It is based on Okapi system with stemmed Index.

prisOpnBasic¶

Run ID: prisOpnBasic
Participant: beijingu-posts-tele.weiran
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: The automatic run, using the title field of the topics, with all opinion-finding features turned off

prisOpnC2¶

Run ID: prisOpnC2
Participant: beijingu-posts-tele.weiran
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: manual
Task: opinion
Run description: Using Indri Query Language to reformulate queries; and using a 2-class(objuective vs subjective) Maximum Entropy classifier to evuluate the polarities of retrievaled documents. This run has an additional run named as prisPolC2 for polarity subtask.

prisOpnC4¶

Run ID: prisOpnC4
Participant: beijingu-posts-tele.weiran
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: manual
Task: opinion
Run description: Using Indri Query Language to reformulate queries; and using a 4-class(scale 1, 2, 3 and 4) Maximum Entropy classifier to evuluate the polarities of retrievaled documents. This run has an additional run named as prisPolC4 for polarity subtask.

prisOpnCm¶

Run ID: prisOpnCm
Participant: beijingu-posts-tele.weiran
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: manual
Task: opinion
Run description: Using Indri Query Language to reformulate queries; and using two Maximum Entropy classifiers(a 4-class one, and a 2-class one) to evuluate the polarities of retrievaled documents. This run has an additional run named as prisPolCm for polarity subtask.

prisPolC2¶

Results | Participants | Input | Summary

Run ID: prisPolC2
Participant: beijingu-posts-tele.weiran
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

prisPolC4¶

Results | Participants | Input | Summary

Run ID: prisPolC4
Participant: beijingu-posts-tele.weiran
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

prisPolCm¶

Results | Participants | Input | Summary

Run ID: prisPolCm
Participant: beijingu-posts-tele.weiran
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

pUB11¶

Results | Participants | Proceedings | Input | Summary

Run ID: pUB11
Participant: ubuffalo.ruiz
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

pUB12¶

Results | Participants | Proceedings | Input | Summary

Run ID: pUB12
Participant: ubuffalo.ruiz
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

pUB21¶

Results | Participants | Proceedings | Input | Summary

Run ID: pUB21
Participant: ubuffalo.ruiz
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

pUB32¶

Results | Participants | Proceedings | Input | Summary

Run ID: pUB32
Participant: ubuffalo.ruiz
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

Relevant¶

Run ID: Relevant
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: An automatic run, using the title fields of the topics, with all opinion-finding features turned off. (Ie a topic-relevance run).

rgu0¶

Run ID: rgu0
Participant: robert-gordonu.mukras
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: All opinion finding features turned off. Simply a Relevance run.

rgu1¶

Run ID: rgu1
Participant: robert-gordonu.mukras
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: All opinion finding features turned on. Edmunds dataset used as background knowledge.

rgu2¶

Run ID: rgu2
Participant: robert-gordonu.mukras
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: All opinion finding features turned on. Rateitall dataset used as background knowledge.

rgu3¶

Run ID: rgu3
Participant: robert-gordonu.mukras
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: All opinion finding features turned on. Scale dataset used as background knowledge.

rgu4¶

Run ID: rgu4
Participant: robert-gordonu.mukras
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: All opinion finding features turned on. Trec dataset used as background knowledge.

SmpStm¶

Run ID: SmpStm
Participant: cas-liu
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: An automatic run, stemming each document and using stactics method.

TDWHU100¶

Results | Participants | Proceedings | Input | Summary

Run ID: TDWHU100
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: This run is based on OKAPI system, with the raw permanlinks part of the BLOGS06 Collection, using only title as queries. Results are ranked based on combinations of doc weights for all relevant feeds.

TDWHU101¶

Results | Participants | Proceedings | Input | Summary

Run ID: TDWHU101
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: This run is based on OKAPI system, with the raw permanlinks part of the BLOGS06 Collection, using the description as queries. Results are ranked based on combinations of doc weights for all relevant feeds.

TDWHU200¶

Results | Participants | Proceedings | Input | Summary

Run ID: TDWHU200
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/18/2007
Type: automatic
Task: feed
Run description: This run is based on OKAPI system, with cleaned permanlinks part of the BLOGS06 Collection, using titles as queries. Results are ranked based on combinations of doc weights for all relevant feeds.

TDWHU201¶

Results | Participants | Proceedings | Input | Summary

Run ID: TDWHU201
Participant: wuhanu.lu
Track: Blog
Year: 2007
Submission: 8/18/2007
Type: automatic
Task: feed
Run description: This run is based on OKAPI system, with cleaned permanlinks part of the BLOGS06 Collection, using the description as queries. Results are ranked based on combinations of doc weights for all relevant feeds.

UALR07Base¶

Run ID: UALR07Base
Participant: uarkansas-littlerock.bayrak
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: base run

UALR07BlogIU¶

Run ID: UALR07BlogIU
Participant: uarkansas-littlerock.bayrak
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: UALR07Base is reranked by the presence of words such as "I", "you", "me", "us", we" and also opinion indicator words such as "like", "feel","think","hate" etc. near the actual query words.

UALR07CML¶

Run ID: UALR07CML
Participant: uarkansas-littlerock.bayrak
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: manual
Task: opinion
Run description: We rerank UALR07Base run by Categorized Machine Learning approach. We use active learning for reranking results by first categorizing the queries into 6 categories namely "event, "location","person","food","thing" and "company".

UALR07IU2¶

Run ID: UALR07IU2
Participant: uarkansas-littlerock.bayrak
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: reranks UALR07TDN base run by identifying the vicinity of words such as "I","you", "me" as well as "like", "feel","think" etc. to the query words

UALR07TDN¶

Run ID: UALR07TDN
Participant: uarkansas-littlerock.bayrak
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: base run with title and description

UALR07Text¶

Run ID: UALR07Text
Participant: uarkansas-littlerock.bayrak
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: We used ad-hoc but intuitive approach to determine opinionated nature of blogs.

uams07bdfreq¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07bdfreq
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Combines topic matching with frequency mathing in a linear combination.

uams07bdtblm¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07bdtblm
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Topic matching with an exponential time-based prior. Feeds selected on highest scoring blog entry.

uams07bdtop¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07bdtop
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Just topic matching, nothing else. Feeds are selected based on their best blog post.

uams07indbl¶

Run ID: uams07indbl
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: Run constructed using Indri (out-of-the-box). No additional features used.

uams07ipolt¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07ipolt
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Task: polarity

uams07mmbl¶

Run ID: uams07mmbl
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: Runs constructed using Mixture Model different indices combined to one. The best mixture, as testen on 2006 data, is 0.15 for title, 0.60 for body and 0.25 for background. No additional features used.

uams07mmq¶

Run ID: uams07mmq
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: Runs constructed using Mixture Model of 0.15 title, 0.60 body and 0.25 background. Query model is constructed using external corpus of AQUAINT-2 (over the same period ad the blog posts) and proper weights are assigned.

uams07mmqcom¶

Run ID: uams07mmqcom
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: Runs constructed using Mixture Model of 0.15 title, 0.60 body and 0.25 background. Query model is constructed using external corpus of AQUAINT-2 (over the same period ad the blog posts) and proper weights are assigned. The number of comments is used as document prior; more comments indicate more discussion and therefore probably a more opinionated post.

uams07mmqop¶

Run ID: uams07mmqop
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: Runs constructed using Mixture Model of 0.15 title, 0.60 body and 0.25 background. Query model is constructed using external corpus of AQUAINT-2 (over the same period ad the blog posts) and proper weights are assigned. A list of opinionated words is used to calculate a ratio of these words for each document and this ratio is added as document prior, giving opinionated posts a higher prior probability.

uams07polbl¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07polbl
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Task: polarity

uams07polq¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07polq
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Task: polarity

uams07polqco¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07polqco
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Task: polarity

uams07polqop¶

Results | Participants | Proceedings | Input | Summary

Run ID: uams07polqop
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Task: polarity

uams07topic¶

Run ID: uams07topic
Participant: uamsterdam.deRijke
Track: Blog
Year: 2007
Submission: 8/7/2007
Type: automatic
Task: opinion
Run description: Runs constructed using Indri. Queries are constructed in every possible combination and are expanded using an external corpus; since blog queries are often news related, we used the AQUAINT-2 corpus for extension. No opinion-features added.

UB1¶

Run ID: UB1
Participant: ubuffalo.ruiz
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: Full query, indri

UB2¶

Run ID: UB2
Participant: ubuffalo.ruiz
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: title with Weibe list

UB3¶

Run ID: UB3
Participant: ubuffalo.ruiz
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: Full query with Weibe

uic1c¶

Run ID: uic1c
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: single classifier, sentence count and IR score combined, full weights for expanded terms

uic1cpnm¶

Results | Participants | Proceedings | Input | Summary

Run ID: uic1cpnm
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

uic1s¶

Run ID: uic1s
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: single classifier, sentence score and IR score combined, full weights for expanded terms

uic1spnm¶

Results | Participants | Proceedings | Input | Summary

Run ID: uic1spnm
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

uic75c¶

Run ID: uic75c
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: single classifier, sentence count score and IR score combined, reduced weights for expanded terms

uic75cpnm¶

Results | Participants | Proceedings | Input | Summary

Run ID: uic75cpnm
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/7/2007
Task: polarity

uic75s¶

Run ID: uic75s
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: single classifier, sentence score and IR score combined, reduced weights for expanded terms

uic75spnm¶

Results | Participants | Proceedings | Input | Summary

Run ID: uic75spnm
Participant: uiuc-zhang
Track: Blog
Year: 2007
Submission: 8/7/2007
Task: polarity

UMaTDPCSwGR¶

Results | Participants | Proceedings | Input | Summary

Run ID: UMaTDPCSwGR
Participant: umass.allan
Track: Blog
Year: 2007
Submission: 8/14/2007
Type: automatic
Task: feed
Run description: unigram language model cluster topic diversity

UMaTiGR¶

Results | Participants | Proceedings | Input | Summary

Run ID: UMaTiGR
Participant: umass.allan
Track: Blog
Year: 2007
Submission: 8/14/2007
Type: automatic
Task: feed
Run description: unigram language model

UMaTiPCS¶

Results | Participants | Proceedings | Input | Summary

Run ID: UMaTiPCS
Participant: umass.allan
Track: Blog
Year: 2007
Submission: 8/14/2007
Type: automatic
Task: feed
Run description: unigram language model cluster

UMaTiPCSwGR¶

Results | Participants | Proceedings | Input | Summary

Run ID: UMaTiPCSwGR
Participant: umass.allan
Track: Blog
Year: 2007
Submission: 8/14/2007
Type: automatic
Task: feed
Run description: unigram language model cluster topic diversity

UniNEblog1¶

Run ID: UniNEblog1
Participant: uneuchatel.savoy
Track: Blog
Year: 2007
Submission: 8/2/2007
Type: automatic
Task: opinion
Run description: Title-only. Indexing using a s-Stemmer, retrieval based on Okapi model.

UniNEblog2¶

Run ID: UniNEblog2
Participant: uneuchatel.savoy
Track: Blog
Year: 2007
Submission: 8/2/2007
Type: automatic
Task: opinion
Run description: Indexing using a s-Stemmer, retrieval based on the Divergence from Randomness Model, PL2. The topic fields used for retrieval are title, description and narrative.

UniNEblog3¶

Run ID: UniNEblog3
Participant: uneuchatel.savoy
Track: Blog
Year: 2007
Submission: 8/2/2007
Type: automatic
Task: opinion
Run description: Data fusion of two IR systems. 1 Retrieval based on Okapi using Rocchio Query-Expansion 2 Retrieval based on Divergence from randomness. The topic fields used for retrieval are title and description.

UniNEblog4¶

Run ID: UniNEblog4
Participant: uneuchatel.savoy
Track: Blog
Year: 2007
Submission: 8/2/2007
Type: automatic
Task: opinion
Run description: Data fusion of two IR systems. 1 Retrieval based on Language Model 2 Retrieval based on Divergence from randomness. The topic field used for retrieval is title

UniNEblog5¶

Run ID: UniNEblog5
Participant: uneuchatel.savoy
Track: Blog
Year: 2007
Submission: 8/2/2007
Type: automatic
Task: opinion
Run description: Data fusion of three IR systems. 1 Retrieval based on Language Model using a 5 gram indexing scheme. 2 Retrieval based on Divergence from randomness. 3. Retrieval using the Okapi model. The topic fields used for retrieval are title and descriptive.

UniNEblog6¶

Run ID: UniNEblog6
Participant: uneuchatel.savoy
Track: Blog
Year: 2007
Submission: 8/2/2007
Type: automatic
Task: opinion
Run description: Indexing using a s-Stemmer, retrieval based on Language model. The topics fields used for retrieval are title and descriptive.

uogBDFeMNZ¶

Results | Participants | Proceedings | Input | Summary

Run ID: uogBDFeMNZ
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Document structure, voting model

uogBDFeMNZhA¶

Results | Participants | Proceedings | Input | Summary

Run ID: uogBDFeMNZhA
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Document structure, voting model, homepages anchortext

uogBDFeMNZP¶

Results | Participants | Proceedings | Input | Summary

Run ID: uogBDFeMNZP
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Document structure, voting model, proximity

uogBDFeMNZpC¶

Results | Participants | Proceedings | Input | Summary

Run ID: uogBDFeMNZpC
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/17/2007
Type: automatic
Task: feed
Run description: Document structure, voting model, cohesiveness

uogBOPF¶

Run ID: uogBOPF
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: A baseline run taking into account document structure

uogBOPFPol¶

Results | Participants | Proceedings | Input | Summary

Run ID: uogBOPFPol
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/6/2007
Task: polarity

uogBOPFProx¶

Run ID: uogBOPFProx
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: A T query run uses document structure evidence and word positions

uogBOPFProxW¶

Run ID: uogBOPFProxW
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run uses T queries with opinion detection.

uogBOPFTD¶

Run ID: uogBOPFTD
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: A TD query run uses document structure evidence

uogBOPFTDOF¶

Run ID: uogBOPFTDOF
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: A TD query run uses document structure evidence and opinion detection

uogBOPFTDW¶

Run ID: uogBOPFTDW
Participant: glasgow.ounis
Track: Blog
Year: 2007
Submission: 8/6/2007
Type: automatic
Task: opinion
Run description: This run uses TD queries with opinion detection.

utblnrr¶

Results | Participants | Proceedings | Input | Summary

Run ID: utblnrr
Participant: utexas-austin.efron
Track: Blog
Year: 2007
Submission: 8/15/2007
Type: automatic
Task: feed
Run description: uses KL divergence between query and feed language models for doc ranking, with feed models calculated by term frequency in the entire text of a feed

utblrr¶

Results | Participants | Proceedings | Input | Summary

Run ID: utblrr
Participant: utexas-austin.efron
Track: Blog
Year: 2007
Submission: 8/15/2007
Type: automatic
Task: feed
Run description: uses KL divergence between query and feed language models for doc ranking, with feed models calculated by term frequency in the entire text of a feed. before presentation, feeds are re-ranked by the likelihood of seeing a query term in a constituent post.

utlc¶

Results | Participants | Proceedings | Input | Summary

Run ID: utlc
Participant: utexas-austin.efron
Track: Blog
Year: 2007
Submission: 8/15/2007
Type: automatic
Task: feed
Run description: uses a linear combination of two methods for computing language models of feeds.

utplmu100¶

Results | Participants | Proceedings | Input | Summary

Run ID: utplmu100
Participant: utexas-austin.efron
Track: Blog
Year: 2007
Submission: 8/15/2007
Type: automatic
Task: feed
Run description: uses KL divergence between query and feed language models for doc ranking, with feed models calculated by term/post frequency

UWbasePhrase¶

Run ID: UWbasePhrase
Participant: uwaterloo-olga
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Baseline run using phrases extracted from the titles. BM25 was used as a document retrieval function.

UWbaseTerms¶

Run ID: UWbaseTerms
Participant: uwaterloo-olga
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Baseline run using single terms extracted from the titles. BM25 was used as a document retrieval function.

UWopinion1¶

Run ID: UWopinion1
Participant: uwaterloo-olga
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Single terms from the titles were used to retrieve the initial document sets. Subjective adjectives co-occurring with the query terms were used in document re-ranking.

UWopinion2¶

Run ID: UWopinion2
Participant: uwaterloo-olga
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Single terms from the topic titles were used to retrieve the initial document sets. Subjective adjectives co-occurring with the phrases from the topic titles were used in document re-ranking.

UWopinion3¶

Run ID: UWopinion3
Participant: uwaterloo-olga
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Phrases from the topic titles were used to retrieve the initial document sets. Subjective adjectives co-occurring with the terms from the topic titles were used in document re-ranking.

UWopinion4¶

Run ID: UWopinion4
Participant: uwaterloo-olga
Track: Blog
Year: 2007
Submission: 8/5/2007
Type: automatic
Task: opinion
Run description: Phrases from the topic titles were used to retrieve the initial document sets. Subjective adjectives co-occurring with the phrases from the topic titles were used in document re-ranking.