Runs - Microblog 2013¶
Avgrank¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: Avgrank
- Participant: Foreseer
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: manual
- Task: adhoc
- MD5:
f9f2b4a249d3852299770f893e7b55bb
- Run description: Active learning. Ranked by combination of SVM and Dirichlet model.
BAUENGFLT¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: BAUENGFLT
- Participant: BAU
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
3ae5f5a499542ed3e22c58a4db53d64f
- Run description: in this run did not used external resources
BAUENGPGRNK¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: BAUENGPGRNK
- Participant: BAU
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
144b693b575f60a641a6d2e18bc3be41
- Run description: in this run did not used external resources
BAUENGSTAT¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: BAUENGSTAT
- Participant: BAU
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
1eb1f56625947714efa56a9889e595bd
- Run description: in this run did not used external resources
BAUENPRKST¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: BAUENPRKST
- Participant: BAU
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
3ffd2363216139b74d9b9d199a31adfc
- Run description: in this run did not used external resources
BJUTCnor¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: BJUTCnor
- Participant: BJUT
- Track: Microblog
- Year: 2013
- Submission: 8/12/2013
- Type: automatic
- Task: adhoc
- MD5:
b555ed468c2f57c5ffe968bf4a334682
- Run description: run3
BJUTEntr¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: BJUTEntr
- Participant: BJUT
- Track: Microblog
- Year: 2013
- Submission: 8/12/2013
- Type: automatic
- Task: adhoc
- MD5:
b33cd4e3cb48f80bf20eff7caf0d0cd0
- Run description: run2
BJUTFreq¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: BJUTFreq
- Participant: BJUT
- Track: Microblog
- Year: 2013
- Submission: 8/12/2013
- Type: automatic
- Task: adhoc
- MD5:
48c44b76748563a15e6e7209c4fca082
- Run description: run1
BNTSrK¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: BNTSrK
- Participant: IRIT
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
2235d94bd74c349450ff628b7eda03ec
- Run description: Bayesian network model for tweet search, topical feature-only
BNTSrKSO¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: BNTSrKSO
- Participant: IRIT
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
888003d4700d1286b9c8495c1fa4bcc5
- Run description: Bayesian network model for tweet search, topical feature, social feature, temporal feature
CIRGIRDISCO2¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: CIRGIRDISCO2
- Participant: CIRG_IRDISCO
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
90ebea3da619f379565134e7641374b8
- Run description: This run uses Wikipedia information for query term expansion by understanding peoples' names specially and rest of information generally.
CIRGIRDISCO3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: CIRGIRDISCO3
- Participant: CIRG_IRDISCO
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
d03542a72075111d517169c1b30b579b
- Run description: This run uses Wikipedia for query term expansion without distinguishing between entity (i.e. peoples' names) and general information.
CIRGIRDISCO4¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: CIRGIRDISCO4
- Participant: CIRG_IRDISCO
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
3ac12fe452496635397ad4968145c096
- Run description: This run utilizes strategy of biased PageRank through text information where node is term and edge is cooccurence relationship between terms.
DFRBase¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: DFRBase
- Participant: UoG_TwTeam
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
8717a055076b3e866cbc47d053faa856
- Run description: This is our DFR baseline
Direrank¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: Direrank
- Participant: Foreseer
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: manual
- Task: adhoc
- MD5:
e2aab98e4d2a2d90d2b08eae8ab1c9f7
- Run description: Active learning. Ranked by Dirichlet model.
FSsvm¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: FSsvm
- Participant: Foreseer
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: manual
- Task: adhoc
- MD5:
b31167dcb6e6166458d2102e5cec3430
- Run description: Active learning. Ranked by SVM.
GSAA¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: GSAA
- Participant: ISIKol
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
8ae64be2788f7c87f04ac75b3aa5721a
- Run description: Title and snippet of Google Search API used for query expansion.
GSAS¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: GSAS
- Participant: ISIKol
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
8a2baa245dff22b85269fe8cdd525598
- Run description: Snippet of Google Search API used for query expansion.
GSAT¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: GSAT
- Participant: ISIKol
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
bd983a4c12d2153bf31c72e6caa0f40b
- Run description: Title of Google Search API used for query expansion.
ICTNETBASE¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: ICTNETBASE
- Participant: ICTNET
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
011f2158271df655a5eb83d9c7f4e36e
- Run description: we simply apply the Bo1 method to expand the queries, and the expanded queries are submitted, and retrieved tweets and their scores are included in this run.
ICTNETBO1EXP¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: ICTNETBO1EXP
- Participant: ICTNET
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
df53ffffa779faf9ba1d9afac94665b5
- Run description: We applied the SVM Rank framework as we did in last year, and a new query expansion method - Bo1, is utilized. A bunch of features are calculated according to the original query and expansion respectively.
ICTNETCOCCUR¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: ICTNETCOCCUR
- Participant: ICTNET
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
96858e5368e4914532eddcf245e1ca45
- Run description: The same method is used as ICTNETBO1EXP, except that word co-occurrence is taken into account while expanding the queries.
ILPSdf¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: ILPSdf
- Participant: UvAILPS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
faab13f520f1e7f59dfb40ab601101ce
- Run description: We use a data fusion method for merging ranked lists returned from several retrieval models.
ILPSl2rB¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: ILPSl2rB
- Participant: UvAILPS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
23aec1a93d9873151d0e91fcd12ec142
- Run description: Baseline run: learning to rank on 15 features extracted from results from API.
ILPSl2rE¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: ILPSl2rE
- Participant: UvAILPS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
c2b6fd8de708eb6f6835e719877ae7b5
- Run description: Identify concepts in queries using Wikipedia, expand query based on results of the concept query on API. Perform learning to rank on results of expanded query.
ILPSub¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: ILPSub
- Participant: UvAILPS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
e661658bb66b58fed03760fa15e15a4c
- Run description: Concepts in query identified using Wikipedia. Query expansion on results of concept query. RSV is doubled if tweet has a URL.
iritfdUrl¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: iritfdUrl
- Participant: IRIT
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
d26c09db969740729eefaa35a2d23885
- Run description: Vector space model+ Url content+ spell checker
iritfdUrlRoc¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: iritfdUrlRoc
- Participant: IRIT
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
091b5e918fe5dc67165b11930aed1ce2
- Run description: Vector space model + Rochhio query expansion+ Url content+ spell checker
kobeRMU¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: kobeRMU
- Participant: KobeU
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
7696b8910df21a572531ca11bd11c852
- Run description: Query likelihood model + word tracking (pseudo-relevance feedback) + URL filtering We filtered out retweets and non-english tweets with a language detection tool called ldig. For word counting, we used external tweets sampled between February 1st and March 31st, 2013.
kobeTSFRM¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: kobeTSFRM
- Participant: KobeU
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: manual
- Task: adhoc
- MD5:
6a39fb9f3fd9426533c4736076cd1e4d
- Run description: Query likelihood model + tweet selection feedback (relevance feedback) + concept tracking (pseudo-relevance feedback). We filtered out retweets and non-english tweets with a language detection tool called ldig. For concept counting, we used external tweets sampled between February 1st and March 31st, 2013.
kobeTSFRMU¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: kobeTSFRMU
- Participant: KobeU
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
035be9a720c0cd1853c919d557116339
- Run description: Query likelihood model + tweet selection feedback (relevance feedback) + concept tracking (pseudo-relevance feedback) + URL filtering. We filtered out retweets and non-english tweets with a language detection tool called ldig. For concept counting, we used external tweets sampled between February 1st and March 31st, 2013.
kobeU¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: kobeU
- Participant: KobeU
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
3948256aff96bc97ebcfc9da6708a8d4
- Run description: Query language model (default) + URL filtering.
ModelSEL922¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: ModelSEL922
- Participant: UoG_TwTeam
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
cf2922f1a2362b78d8315ffcabb11c1d
- Run description: Model selection approach. We select an optimal retrieval model, based on pre-retrieval features
NOVAsearch00¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: NOVAsearch00
- Participant: NOVASEARCH
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
56794c7d892389aebb57997d8e12de70
- Run description: This run performs duplicates, retweet and language filtering.
NOVAsearch01¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: NOVAsearch01
- Participant: NOVASEARCH
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
8d65a973b60586fa7ba0f230343a4d5f
- Run description: Using temporal reranking with respect to the time of the query.
NOVAsearch02¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: NOVAsearch02
- Participant: NOVASEARCH
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
b93fd56d9657b3503f392edc2cf4e22e
- Run description: Temporal reranking using Wikipedia pageview statistics for the top page resulting from the search of the query on Wikipedia
NOVAsearch03¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: NOVAsearch03
- Participant: NOVASEARCH
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
e7237b7fb585134b816419b90106cbdd
- Run description: Makes use of a crawl of documents linked from tweets done in 2013-08-14. Only the titles are used.
PKUICST1¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PKUICST1
- Participant: PKUICST
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
3b0e914a2880560b7ca38192071fd7ff
- Run description: Using a local corpus of tweets collected via the official Twitter streaming API over a two-month period 1 February, 2013 - 31 March. Those resources are timely with respect to the queries Using Ranking SVM with different semantic features to train a model. Features are Language Model and BM25 scores on different indexes (Origin and DocEx corpus). Training on 59 queries and their related official labeled tweets released by TREC2012.
PKUICST2¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PKUICST2
- Participant: PKUICST
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
b2fd3fdd6223f66bfe2099bdd61e0cf6
- Run description: Using a local corpus of tweets collected via the official Twitter streaming API over a two-month period 1 February, 2013 - 31 March. Those resources are timely with respect to the queries Using BM25 Model, Query Expanded with the highest score tweet. Run on DocEx Corpus (Origin Corpus + Link Info Corpus).
PKUICST3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PKUICST3
- Participant: PKUICST
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
784cfa27f2a828d0d96a7a908c8c5b4f
- Run description: Using a local corpus of tweets collected via the official Twitter streaming API over a two-month period 1 February, 2013 - 31 March. Those resources are timely with respect to the queries Using Language Model, Query Expanded with the highest score tweet. Run on DocEx Corpus (Origin Corpus + Link Info Corpus).
PKUICST4¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PKUICST4
- Participant: PKUICST
- Track: Microblog
- Year: 2013
- Submission: 8/13/2013
- Type: automatic
- Task: adhoc
- MD5:
60e3b4f6125f3dc89bd0c777f27aa7b4
- Run description: An internal run relies strictly on data obtained from the API. Using Language Model and Query Expanded with the highest score tweet from baseline language model. Results are filtered based on some simple rules.
PrisRun1¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PrisRun1
- Participant: PRIS
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
3cac9e45dab154ced788363018b0c31e
- Run description: Only use API result,without any other external resources.
PrisRun2¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PrisRun2
- Participant: PRIS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
0ad0fe4124e0c3b41a107837d7dc18c7
- Run description: using API and url information
PrisRun3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PrisRun3
- Participant: PRIS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: manual
- Task: adhoc
- MD5:
52b3c6b8c4e272808524d18fb7e9853f
- Run description: use expansion words and URL!
PrisRun4¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: PrisRun4
- Participant: PRIS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
4ea8dceb4bca8153c5785be0a56b87d4
- Run description: use external URL information!
QCRI1¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QCRI1
- Participant: QCRI
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
dc23ccffd5a94f6844cd06f32f0e2348
- Run description: single learning to rank model
QCRI2¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QCRI2
- Participant: QCRI
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
50540c34bbf8e0de8afcc54b41df001e
- Run description: single learning to rank model tuned with selected validated dataset.
QCRI3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QCRI3
- Participant: QCRI
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
72f8e54ca134836424c865a4d3b57d2c
- Run description: Ensemble 12 learning to rank model
QCRI4¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QCRI4
- Participant: QCRI
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
396093a1807a5ca5abbf21733679a8a9
- Run description: Ensemble 12 learning to rank model incorporating external information. Use google for query expansion.
QEClustIDF¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QEClustIDF
- Participant: UoG_TwTeam
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
be1dc9b77283103ed9b5b6afb0f4223c
- Run description: We performed clustering over the collection, previous to retrieval. Those clusters were used in a separate index, as evidence to extract expansion terms. We take care to use information previous to the query time. Terms in clusters are then ranked by a discounted IDF function.
QEDiscIDF25Good¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QEDiscIDF25Good
- Participant: UoG_TwTeam
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
003d7f3a4d23e4a28f15e8eb0d1047d4
- Run description: Model selection approach, with discounted IDF scoring of PRF terms
QUBaseline¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QUBaseline
- Participant: QU
- Track: Microblog
- Year: 2013
- Submission: 8/8/2013
- Type: automatic
- Task: adhoc
- MD5:
7b4d5e5262800e641933413cbe932af5
- Run description: The run utilizes query expansion to expand each topic and provide search results based on the expanded topic. This is achieved by submitting the original query to the search client and then use the retrieved results to expand the query. In addition, retweets removal is applied to eliminate retweets from the results.
QUDocExp¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QUDocExp
- Participant: QU
- Track: Microblog
- Year: 2013
- Submission: 8/8/2013
- Type: automatic
- Task: adhoc
- MD5:
59e0d354239312b23e1327b2784e0c19
- Run description: The run utilizes document expansion to expand each document in the results list of a query and rescore results. The expansion considers the lexical and temporal aspects of a tweet. In addition, retweets removal is applied to eliminate retweets from the results. As for the external resources, a third party language detection tool is used to eliminate non-English tweets. The tool is timely with respect to the queries.
QUQueryExp¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QUQueryExp
- Participant: QU
- Track: Microblog
- Year: 2013
- Submission: 8/8/2013
- Type: automatic
- Task: adhoc
- MD5:
1e786c614de86b98bbe0de2ed8a6a94e
- Run description: The run utilizes query expansion to expand each topic and provide search results based on the expanded topic. This is achieved by submitting the original query to the search client and then use the retrieved results to expand the query. In addition, retweets removal is applied to eliminate retweets from the results. As for the external resources, a third party language detection tool is used to eliminate non-English tweets. The tool is timely with respect to the queries.
QUTemporal¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: QUTemporal
- Participant: QU
- Track: Microblog
- Year: 2013
- Submission: 8/8/2013
- Type: automatic
- Task: adhoc
- MD5:
e54141509be5047db67788a472691f70
- Run description: For each query, the results retrieved through the search API are rescored based on their tweet time in relation with the the query tweet time. In addition, retweets removal is applied to eliminate retweets from the results. As for the external resources, a third party language detection tool is used to eliminate non-English tweets. The tool is timely with respect to the queries.
RvsDir¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: RvsDir
- Participant: Foreseer
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: manual
- Task: adhoc
- MD5:
f2f5e774b71539bedff10924c051e5c0
- Run description: Active learning. Ranked by combination of SVM and reversed Dirichlet model.
scunce1¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: scunce1
- Participant: scunce
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
a6cf9c798ff85c5deb59c268a879cbe6
- Run description: feature1: weighted words for query expansion and get each tweets value. feature2: KL value from original query and tweets feature3: KL value from expanded query(top 15 tweets) and tweets
scunce2¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: scunce2
- Participant: scunce
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
5defb5b47e354fbc4e26daffa88bd610
- Run description: feature1: weighted words for query expansion and get each tweets value. feature2: KL value from original query and tweets feature3: KL value from expanded query(top 15 tweets) and tweets use 2011 and 2012 topics for train
scunce3¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: scunce3
- Participant: scunce
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
d6d48cc6727152e6dfe1128811cd648a
- Run description: feature1: weighted words for query expansion and get each tweets value. feature2: KL value from original query and tweets feature3: KL value from expanded query(top 15 tweets) and tweets feature4: weighted words for query expansion using google search(top30) and get each tweets value. feature5: KL value from expanded query(top20 google search title) and tweets use 2011 topics for train
scunce4¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: scunce4
- Participant: scunce
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
c620e70083ddf6533588012a4eaf0c99
- Run description: feature1: weighted words for query expansion and get each tweets value. feature2: KL value from original query and tweets feature3: KL value from expanded query(top 15 tweets) and tweets feature4: weighted words for query expansion using google search(top30) and get each tweets value. feature5: KL value from expanded query(top20 google search title) and tweets use 2011 and 2012 topics for train
stan1kl¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: stan1kl
- Participant: scwhu
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
1c247ad0f01cf9ee19f2a9bf9672e0c2
- Run description: Use pseudo feedback to do query expand(top7) and KL-divergence to rank.
stan2kl¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: stan2kl
- Participant: scwhu
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
00512d87119f1d3cf77ba130621398c7
- Run description: Use pseudo feedback to do query expand(top7).
stan3kl¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: stan3kl
- Participant: scwhu
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
5b8a91472de37685e4d8a746543adcd1
- Run description: Use pseudo feedback to do query expand(top10) and KL-divergence to rank.
stan4col¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: stan4col
- Participant: scwhu
- Track: Microblog
- Year: 2013
- Submission: 8/15/2013
- Type: automatic
- Task: adhoc
- MD5:
93f4085e5285bb6799bf40019ac4365a
- Run description: Use pseudo feedback and word collocation to do query expand.
UCASgem¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UCASgem
- Participant: UCAS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
7a558df543395a4a2878d2845a8b980d
- Run description: Noting specifical
UCASqe¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UCASqe
- Participant: UCAS
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
c035bade7a98da59687c730f13b222c6
- Run description: Noting specifical resources
UDInfoMB¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UDInfoMB
- Participant: udel_fang
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
b52dc381e774c94e5ad894a3a41de01c
- Run description: Initial retrieval from the API used query suggestion from Yahoo! Used lang_id.py as the language identification tool.
UDInfoMINT¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UDInfoMINT
- Participant: udel_fang
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
9af9e5dec2a1e303031256615eed8b5c
- Run description: Used only internal information. Index is built based on initial run with original query and per query term.
UDInfoMTB1¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UDInfoMTB1
- Participant: udel_fang
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
808a4ef220392e1349e2cbd3b2291dd9
- Run description: Initial retrieval from the API used query suggestion from Yahoo! Used lang_id.py as the language identification tool.
UDInfoMTB2¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UDInfoMTB2
- Participant: udel_fang
- Track: Microblog
- Year: 2013
- Submission: 8/14/2013
- Type: automatic
- Task: adhoc
- MD5:
97506f9bf0a3cf4538c05807c110b6c3
- Run description: Initial retrieval from the API used query suggestion from Yahoo! Used lang_id.py as the language identification tool.
WISSySeCo¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: WISSySeCo
- Participant: wistud
- Track: Microblog
- Year: 2013
- Submission: 8/16/2013
- Type: automatic
- Task: adhoc
- MD5:
33461b19c26d2062637a7cc13986037a
- Run description: Used NER tools to identify the semantics in the tweets