Runs - Medical 2012¶

AEHRClvl0¶

Run ID: AEHRClvl0
Participant: AEHRC
Track: Medical
Year: 2012
Submission: 8/5/2012
Type: automatic
Task: main
MD5: 19042f7231165e1623d1bc6c0a7c8641
Run description: Graph-based inference model - level 0.

AEHRClvl1¶

Run ID: AEHRClvl1
Participant: AEHRC
Track: Medical
Year: 2012
Submission: 8/5/2012
Type: automatic
Task: main
MD5: 5dcdb0b1dc8f744c2abb4b927d35bcfe
Run description: Graph-based inference model - level 1.

AEHRClvl2¶

Run ID: AEHRClvl2
Participant: AEHRC
Track: Medical
Year: 2012
Submission: 8/5/2012
Type: automatic
Task: main
MD5: d34165fb8485dbed6713d8197bba4ed2
Run description: Graph-based inference model - level 2.

AEHRCsub¶

Run ID: AEHRCsub
Participant: AEHRC
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 65a5e1e438a4c843ba673ea7fa8a15a0
Run description: AEHRC run based on concepts and subsumptions (SNOMED CT).

APRel1¶

Run ID: APRel1
Participant: RMIT
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 3d4601e3a3256763bc4779423a514e9e
Run description: This run uses negex algorithm to spot negated concepts before the indexing stage and extends icd codes in the documents. It is mainly based on pseudo relevance feedback, and exploits an automatically generated list of relevant documents in the collection that have icd codes from the queries. Combined with Rocchio for query expansion using BO1 model from terrier open source search engine.

APRel2¶

Run ID: APRel2
Participant: RMIT
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 54695b2b5820ffad4427d083e45ff814
Run description: Thir run is mostly similar to RMIT's first run with a difference in the way pseudo relevant documents were gathered. In this run we use two most relevant ICD codes for each query as evidence to mark relevant documents. Ranking Model used is PL2 model for retrieval and BO1 for query expansion from Terrier's open source search engine.

atigeo0¶

Run ID: atigeo0
Participant: xMusketeers
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 34aa474fe5e0cf9287f5cd92f7c2a62d
Run description: Baseline system with no ICD9 code descriptions

atigeo1¶

Run ID: atigeo1
Participant: xMusketeers
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 4372205201bd2d97dfb5fb792ea1b97e
Run description: Baseline system with minimal ICD9 code descriptions

atigeo2¶

Run ID: atigeo2
Participant: xMusketeers
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: fee9b52bbb0c12a6a3a014d8b4be8500
Run description: Baseline system with moderate ICD9 code descriptions

atigeo3¶

Run ID: atigeo3
Participant: xMusketeers
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 3236d4395135451bbb398407c92120b0
Run description: Baseline system with maximum ICD9 code descriptions

BMIUOUbase¶

Run ID: BMIUOUbase
Participant: BMIUOU
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: fc61cf3d80ebb1cade9a4199403d4720
Run description: Use Lucene for query without expansion.

BMIUOUens¶

Run ID: BMIUOUens
Participant: BMIUOU
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 3d3da4a85da4e10f4c1873b12a07db97
Run description: Based on Lucene, we do query expansion combining 3 methods synonym expansion, predication expansion, and relation expansion using UMLS. Both expanded and non-expanded words are weighted by idf.

BMIUOUensneg¶

Run ID: BMIUOUensneg
Participant: BMIUOU
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 6fccffbeb2aee34b53deffe5e5928963
Run description: Based on Lucene, we do query expansion combining 3 methods synonym expansion, predication expansion, and relation expansion using UMLS. Both expanded and non-expanded words are weighted by idf. We reduced weight for documents that have negation term.

BMIUOUsyn¶

Run ID: BMIUOUsyn
Participant: BMIUOU
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 5fb8e94616af88c1cd50b576348053c7
Run description: This method used a query expansion based on UMLS and the Lucene search engine. We weight non-expanded words according to idf in the corpus.

buptprisBase¶

Run ID: buptprisBase
Participant: PRIS
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 86049d42abc2294b2206448140f725f7
Run description: A baseline using Lucene,with query expanded by several tools including MetaMap,UMLS Metathesaurus and SNOMED,and ICD9 information mining.The weight of each indexed field is defined by personal experiences.Result scores computed with lucene retrival scores.

buptprisCscr¶

Run ID: buptprisCscr
Participant: PRIS
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: b592b5700e27677324cbaf855987bd9c
Run description: The buptprisCscore run considers "contents" field exclusively when getting the final score of each returned visit.With intersection algorithm and few-result-deal algorithm the same as buptprisInt.

buptprisInt¶

Run ID: buptprisInt
Participant: PRIS
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 6d8cb14948b0352aca4d082ad9d201b0
Run description: To make improvement to buptpris_Base,this run split a query into several subquerys,and compute the intersection of their retrival results.At the same time,this run include a algorithm to deal with the topics returning few results.

buptprisLrnk¶

Run ID: buptprisLrnk
Participant: PRIS
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 7aabaae3296544267d3dea80797a73ba
Run description: A try to improve the ranking with learning to rank algorithm on the basis of buptpris_Int run.

DCU21¶

Run ID: DCU21
Participant: DCU
Track: Medical
Year: 2012
Submission: 8/1/2012
Type: automatic
Task: main
MD5: ebc5858309265edceeb507430d51b9a2
Run description: standard BM25 with default parameters

DCU22¶

Run ID: DCU22
Participant: DCU
Track: Medical
Year: 2012
Submission: 8/1/2012
Type: automatic
Task: main
MD5: 432098e9bb582d3c563bc6533d794e1a
Run description: standard BM25 with default parameters, PRF with 10 docs, 10 terms, filtering

DCU23b¶

Run ID: DCU23b
Participant: DCU
Track: Medical
Year: 2012
Submission: 8/1/2012
Type: automatic
Task: main
MD5: f31dd0f8eb61f5b1441fe0778e1fa154
Run description: standard BM25 with default parameters on concept-expanded query

DCU24b¶

Run ID: DCU24b
Participant: DCU
Track: Medical
Year: 2012
Submission: 8/1/2012
Type: automatic
Task: main
MD5: 231dd72ffdb2d8cc4ebbc186ac93c013
Run description: standard BM25 with default parameters on concept-expanded query, query expansion using 10 terms, 10 docs, filtering

EssieAuto¶

Run ID: EssieAuto
Participant: NLM
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 638f7722d9a9ed6a5e262088697bbae3
Run description: Essie search using frames over positive sections padded with lossy expansion

GE4¶

Run ID: GE4
Participant: RMIT
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 522e349225ca8ae31e3bb65fd7268f29
Run description: This run uses negex algorithm to spot negated concepts before the indexing stage and extends icd codes in the documents. Using 2 external knowledge sources, namely: UMLS and Dbpedia it expands the extracted concepts from the metamap output and based on the reliability of the semantic types (previsouly trained) from the UMLS, it selects the best candidates for expansion.

ikmlab¶

Run ID: ikmlab
Participant: IKMLAB
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 34223ed8f516c7a1d7b2fbc17c03f5ad
Run description: We used the UMLS concept to expansion topics

ikmlab2¶

Run ID: ikmlab2
Participant: IKMLAB
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: ed77b80ea254476e3e31b6621e4d8443
Run description: UMLS concept and bigram used

LSIS1¶

Run ID: LSIS1
Participant: LSIS
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 3e86967f890051f8679040e57e2e1c20
Run description: This run uses MetaMap for medical corpus conceptualization with added best concept strategy, after that the model In_expc2 based on DFR is used for matching the documents and the conceptualized queries by MetaMap with added best concept strategy and expanded using Bo1 model.

LSIS2¶

Run ID: LSIS2
Participant: LSIS
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 77dfe50107d0ded8a7716afeb7f2660c
Run description: This run uses MetaMap for medical corpus conceptualization with added best concept strategy, after that the model In_expc2 based on DFR is used for matching the documents and the expanded queries using Bo1 model.

LSIS3¶

Run ID: LSIS3
Participant: LSIS
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 8bcee7a63143cd4f261eeddc08e05986
Run description: This run uses the original medical corpus, the model In_expc2 based on DFR is used for matching the documents and the expanded queries using Bo1 model.

MayoExpanded¶

Run ID: MayoExpanded
Participant: MayoClinicNLP
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 34446a19e94492d3395548d22ee4affe
Run description: Queries in this run are expanded via the UMLS and via k-Nearest Neighbor random-indexed semantic vectors (which are trained on millions of Mayo Clinic notes). Query terms are weighted based on the number of synonyms containing that term.

MayoLucene¶

Run ID: MayoLucene
Participant: MayoClinicNLP
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: fcfa2a94e0c06d1349434666dca3d766
Run description: A baseline system using Apache Lucene.

MayoMetaData¶

Run ID: MayoMetaData
Participant: MayoClinicNLP
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: fd94fa57e7a08092efed9c15a245368e
Run description: An automatic run using Apache Lucene, testing the value of structured data in information retrieval. Metadata (including ICD-9 codes from admit and discharge diagnoses) are additionally converted into text and searched.

MayoPayload¶

Run ID: MayoPayload
Participant: MayoClinicNLP
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 6e65158bf617d2f25d7f9e4fe3b5c7da
Run description: An automatic run using Apache Lucene, testing the value of using traditional NLP/Information Extraction output in clinical information retrieval. Specifically, Mayo Clinic's MedTagger pipeline was used to process the reports, and these results were stored in a Lucene index via payloads.

NICTAUBC1¶

Run ID: NICTAUBC1
Participant: NICTA
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 57dda8902954aa873ed9ca916e556f67
Run description: This system is based in the combination of two expansion techniques: graph-based concept similarity ranking, and hand-crafted rules based on semantic types and DBpedia. The query is first parsed with MetaMap and Negex, and the Personalised Pagerank technique is used to obtain a ranking of concepts related to the original query. The query is further expanded by adding DBpedia terms that are related to the target concepts (and fall under a list of predefined UMLS semantic types).

NICTAUBC2¶

Run ID: NICTAUBC2
Participant: NICTA
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: e9e3672c21f04049617a289d3e1cf9ac
Run description: A variant of NICTAUBC1, in this case the DBpedia expansion occurs both before and after the Pagerank process. The indexing and the pagerank thesholds change slightly (reports and no stemming for the indexing; threshold of 4 pagerank expansion terms per query).

NICTAUBC4¶

Run ID: NICTAUBC4
Participant: NICTA
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 090f7eb7c15ed23f50ddb0652a9f1839
Run description: A variant of NICTAUBC1, in this case the Personalised Pagerank expansion is omitted, and the expansion terms are obtained from DBpedia.

NICTAUBC6¶

Run ID: NICTAUBC6
Participant: NICTA
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 8668836b4a275a41ccc92cf27b057965
Run description: This system is a variant of BCUNICTA1, and the main difference is the sequence of application of the expansion techniques. First the initial query is expanded using DBpedia and UMLS, and then the Personalised Pagerank process is executed to add further concepts.

NLMLuceneExp¶

Run ID: NLMLuceneExp
Participant: NLM
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: e8d9b4fff62e3510e9926fc47b4039ac
Run description: NLMLuceneExp - It uses as well the PICO frames to identify relevant search terms and to do expansion of drugs and procedures. Resources---Wikipedia and Google search

NLMLuceneSec¶

Run ID: NLMLuceneSec
Participant: NLM
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: c96d4e428c73fb9c3ed212359af00b7d
Run description: NLMLuceneSec - It uses as well the PICO frames to identify relevant search terms and to do expansion of drugs and procedures. Search terms are weighted according to the section of the report in which they appear. Resources---Wikipedia and Google

NLMManual¶

Run ID: NLMManual
Participant: NLM
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: manual
Task: main
MD5: 4e039ed68707734eafa2b1a5cea10154
Run description: Interactively refined queries padded with an automatic run based on frames search over positive text in sections

OHSUCEtICD¶

Run ID: OHSUCEtICD
Participant: OHSU
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 9158f6080e940a69670b4d26fb8efe95
Run description: Automatically-generated queries, using both topic words as well as MetaMap preferred terms, along with alternate terms derived from MeSH, as well as automatically-assigned ICD codes.

OHSUCombET¶

Run ID: OHSUCombET
Participant: OHSU
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: c9ac27a1dd36b63afde04b456966fc27
Run description: Automatically-generated queries, using both topic words as well as MetaMap preferred terms, along with alternate terms derived from MeSH.

OHSUCombICD¶

Run ID: OHSUCombICD
Participant: OHSU
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 780d3e9d50bbcc1fb9255857db8a00f8
Run description: Automatically-generated queries, using both topic words as well as MetaMap preferred terms, along with automatically-assigned ICD codes.

ohsuManBool¶

Run ID: ohsuManBool
Participant: OHSU
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: manual
Task: main
MD5: a15c04a2cdfa364adbd6b85802921e2a
Run description: Manually-generated query using Boolean features of search system, as well as ICD code filters.

quta¶

Run ID: quta
Participant: qutir12
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 3fff8ebc0245918e390054d48bf07079
Run description: No external data used in this run Topsig run, 32768-bit signatures, using term stats SPLIT-TYPE = hard SPLIT-MAX = 200 TOPIC-OUTPUT-K = 1000 PSEUDO-FEEDBACK-SAMPLE = 0

qutb¶

Run ID: qutb
Participant: qutir12
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 98be81f725ca68704c1d02238ec90876
Run description: Topsig run, 32768-bit signatures, using term stats SPLIT-TYPE = hard SPLIT-MAX = 200 TOPIC-OUTPUT-K = 1000 PSEUDO-FEEDBACK-SAMPLE = 0 Using snomed-ct to perform query culling. Terms or groups of terms that are not descriptions in snomed-ct are removed from the queries automatically.

qutc¶

Run ID: qutc
Participant: qutir12
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: c7d537106cab10c8d651ba1e132fceaa
Run description: Topsig run, 32768-bit signatures, using term stats SPLIT-TYPE = hard SPLIT-MAX = 200 TOPIC-OUTPUT-K = 1000 PSEUDO-FEEDBACK-SAMPLE = 0 Using snomed-ct to perform query culling. Terms or groups of terms that are not descriptions in snomed-ct are removed from the queries automatically. Using improved term stats file (in comparison to runs quta and qutb)

RAPRel2¶

Run ID: RAPRel2
Participant: RMIT
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 52771aeb9c0655f57f24ba0214f12d8c
Run description: This run has the same settings as run-1 however, we use a different technique to create a pseudo relevance qrel to improve our pseudo relevance feedback method. Relevant documents are ranked based on their calculted Idf, taking all ICD codes for each document as terms in the documents. Top 10 documents are selected for each query and used as relevance feedback. This run aims to remove the noise potentially added by our first method.

sennamed1¶

Run ID: sennamed1
Participant: sennamed
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: ddbe794c9cfd261002b9a2b5e4cce74d
Run description: metamap umls concepts + indri language model + (blind) relevance feedback

sennamed2¶

Run ID: sennamed2
Participant: sennamed
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 514f2ae80a8742c4d0620773c1fab4e7
Run description: metamap umls concepts + simple tfidf distances + (blind) relevance feedback

sennamed3¶

Run ID: sennamed3
Participant: sennamed
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: 2efb601a32a7ab4e9a4aa1632828303a
Run description: metamap umls concepts + lemur-based tfidf

sennamedlsi¶

Run ID: sennamedlsi
Participant: sennamed
Track: Medical
Year: 2012
Submission: 8/9/2012
Type: automatic
Task: main
MD5: dff41aa3eed16dfc23bec4cbbc848b86
Run description: metamap umls concept extraction + LSI

Siena1¶

Run ID: Siena1
Participant: SCIAITeam
Track: Medical
Year: 2012
Submission: 8/6/2012
Type: automatic
Task: main
MD5: 27906149be25eed24011b7835eadc46a
Run description: Lucene Negation Tagger Stanford Post Tagger ICD Codes

Siena2¶

Run ID: Siena2
Participant: SCIAITeam
Track: Medical
Year: 2012
Submission: 8/6/2012
Type: automatic
Task: main
MD5: 3d726b6ff761a031c8ff66d53153a8b0
Run description: Lucene Negation Tagger Stanford Post Tagger ICD Codes UMLS Expansion

Siena3¶

Run ID: Siena3
Participant: SCIAITeam
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: manual
Task: main
MD5: fb136d7c37587f989f536f6f98d05f69
Run description: Matrix Run Had to remove lines after printed out.

SNUBME1¶

Run ID: SNUBME1
Participant: SNU_BME
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 108e33a5c122f31564e2414270f711e8
Run description: run Indri with language model(dirichlet)

SNUBME2¶

Run ID: SNUBME2
Participant: SNU_BME
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 9196a871275c007f653970af0f404adb
Run description: Indri(run method---two_default) no external resources.

SNUBME3¶

Run ID: SNUBME3
Participant: SNU_BME
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 679fde53064c8c5d2aa9d8938d4280ae
Run description: Run Indri with dirichlet method. Add MetaMap score to the result above. external resources : MetaMap

SNUBME4¶

Run ID: SNUBME4
Participant: SNU_BME
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: manual
Task: main
MD5: c7a9b8afbb6f1cb825807ccb3b86bb5e
Run description: manually added query weights. run using Indri with Two-staging

UCDCSI1¶

Run ID: UCDCSI1
Participant: UCD_CSI
Track: Medical
Year: 2012
Submission: 8/2/2012
Type: manual
Task: main
MD5: 4098cef62073c6bb1b4613a841ffe1cc
Run description: A run using MeSH based expansions, Indri's structured query language to specify demographic information such as ages and Concept Re-Ranking.

UCDCSI2¶

Run ID: UCDCSI2
Participant: UCD_CSI
Track: Medical
Year: 2012
Submission: 8/2/2012
Type: manual
Task: main
MD5: aa05d74bcb9747081d842e176ebbef7f
Run description: A run using MeSH based expansions, Indri's structured query language to specify demographic information such as ages and Concept Re-Ranking. Furthermore it uses field-based retrieval in order to utilise more specific information regarding medical conditions namely determining the experiencer and whether or not it occurred in the past.

UCDCSI3¶

Run ID: UCDCSI3
Participant: UCD_CSI
Track: Medical
Year: 2012
Submission: 8/2/2012
Type: automatic
Task: main
MD5: 1c737e595b3f176dca185ba6e65505f1
Run description: An automatic run using MeSH based expansions, Indri's structured query language to specify demographic information such as ages and Concept Re-Ranking.

UCDCSI4¶

Run ID: UCDCSI4
Participant: UCD_CSI
Track: Medical
Year: 2012
Submission: 8/2/2012
Type: automatic
Task: main
MD5: 8898c2e26a7e7be93d6225d51ad1163b
Run description: An automatic run using expansions from the Freebase graph, Indri's structured query language to specify demographic information such as ages and Concept Re-Ranking.

ucm1¶

Run ID: ucm1
Participant: NIL_UCM
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 800cc4f5fd52c12a689b0ee3a5bf16a5
Run description: Lucene, basic analyzer, process by report, score per visit compute as the sum of scores per report associated to visit, hitsPerPage = 100

ucm3¶

Run ID: ucm3
Participant: NIL_UCM
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 9e19f5a93c57f2f34da4d97a2ee64d84
Run description: Lucene, basic analyzer, process by report, score per visit compute as the sum of scores per report associated to visit, hitsPerPage = 1000

ucm4¶

Run ID: ucm4
Participant: NIL_UCM
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 6b2e6c612e71d2d52f907a25ebff104f
Run description: Lucene, basic analyzer, process by report, score per visit compute as the best score per report associated to visit, hitsPerPage = 1000

ucm5¶

Run ID: ucm5
Participant: NIL_UCM
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 4e3e60feccddc9318e352dbd4d13e960
Run description: Lucene, basic analyzer, process by report, score per visit compute as the best score per report associated to visit, hitsPerPage = 1000, index only with reports with visit not NULL

udelMED¶

Run ID: udelMED
Participant: udel
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 821ef9ed8be6bb18f85357a4e6235245
Run description: Less advanced IR language models using external expansions. External expansion collections include ClueWeb09 CatB dataset, 07 Genomics Track dataset, and 2012 MESH Terms.

udelMNZ¶

Run ID: udelMNZ
Participant: udel
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: e2cce9352220fcaee6ae41c5873858bc
Run description: Method: advanced IR statistical model with external expansion External Collections: ClueWeb09 Cat.B, 2007 Genomics Track dataset, MESH 2012

udelMRF¶

Run ID: udelMRF
Participant: udel
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: d467b4d001a2ccd4a756886c12ea4e55
Run description: Advanced IR language models using external expansions. External expansion collections include ClueWeb09 CatB dataset and 07 Genomics Track dataset.

udelSUM¶

Run ID: udelSUM
Participant: udel
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 9a52776d62d70644e31aa4dd5b6fad1b
Run description: Advanced IR language models using external expansions. Expansion collections include ClueWeb09 CatB dataset, 07 Genomics Track dataset, and MESH 2012.

UDInfoMed1¶

Run ID: UDInfoMed1
Participant: udel_fang
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 63efc10f08d4662d9368cd03b6b6b5d5
Run description: This run uses the original terms in queries and reports. Uses the TwoStage rule.

UDInfoMed12¶

Run ID: UDInfoMed12
Participant: udel_fang
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: c717d19e6b1d75b5e3d95d58e854c700
Run description: This run is based on UMLS mapping. The Unified Medical Language System has been used to map all the terms in the medical reports to a specific medical concept code. This expanded all the words and phrases to a equivalent coded language. All native candidates for each term was used.

UDInfoMed123¶

Run ID: UDInfoMed123
Participant: udel_fang
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 499ed7cf9b855964bbcd5ef266f05c21
Run description: This run is based on UMLS mapping. The Unified Medical Language System has been used to map all the terms in the medical reports to a specific medical concept code. This expanded all the words and phrases to a equivalent coded language.

uogTrMConQ¶

Run ID: uogTrMConQ
Participant: uogTr
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 36c908e07f2dff694c86eff7e1704251
Run description: Description---This run combines Divergence from Randomness and Voting Model approaches to medical visit ranking. It applies a novel technique that deals with the medical context of terms in queries and documents, and query expansion.

uogTrMConQRa¶

Run ID: uogTrMConQRa
Participant: uogTr
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: d7228a6adf8d8c32fc757567c91ef6e0
Run description: Description: This run combines Divergence from Randomness and Voting Model approaches to medical visit ranking. It applies a novel technique that deals with the medical context of terms in queries and documents, and query expansion, as well as a novel approach to process hidden medical knowledge.

uogTrMConQRd¶

Run ID: uogTrMConQRd
Participant: uogTr
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 53ee41399f2c5685b070b0d9162bab83
Run description: Description: This run combines Divergence from Randomness and Voting Model approaches to medical visit ranking. It applies a novel technique that deals with the medical context of terms in queries and documents, and query expansion, as well as a novel approach to process specific types of hidden medical knowledge.

uogTrMConQT¶

Run ID: uogTrMConQT
Participant: uogTr
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: c5800c2e9d706d85fa0cff6a54f795fe
Run description: This run combines Divergence from Randomness and Voting Model approaches to medical visit ranking. It applies a novel technique that deals with the medical context of terms in queries and documents, and query expansion, as well as a novel approach to deal with hidden information within the medical records.

USFISDS1¶

Run ID: USFISDS1
Participant: USF_ISDS
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 6e3070b7fef73f48c72e866917565b8d
Run description: All 50 topic run with non function words removed from topic descriptions. Weighting given to hit counts. Only documents with deduped VisitIds

USFISDS2¶

Run ID: USFISDS2
Participant: USF_ISDS
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 0f0cec0afed710f426b34edfca958287
Run description: All 50 topic run with non function words removed from topic descriptions. Weighting given to hit counts. Only documents with deduped VisitIds. documents with the same visitid evaluated together.

UTDHLTA¶

Run ID: UTDHLTA
Participant: UTDHLT
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: a89eb6d20435459442956152b69564af
Run description: BM25, cohort constraint re-ranking, shallow keyword deconstruction, assertions; utilizes PubMed Central, Wikipedia, UMLS, and WordNet.

UTDHLTASK¶

Run ID: UTDHLTASK
Participant: UTDHLT
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 0fcc995622b9a36f9375ddb45446e7db
Run description: BM25, cohort constraint re-ranking, deep keyword deconstruction, assertions; utilizes PubMed Central, Wikipedia, UMLS, and WordNet.

UTDHLTNA¶

Run ID: UTDHLTNA
Participant: UTDHLT
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 479da76e9c1c8065921f01cf5a9be7d9
Run description: BM25, cohort constraint re-ranking, shallow keyword deconstruction, no assertions; utilizes PubMed Central, Wikipedia, UMLS, and WordNet.

UTDHLTNASK¶

Run ID: UTDHLTNASK
Participant: UTDHLT
Track: Medical
Year: 2012
Submission: 8/8/2012
Type: automatic
Task: main
MD5: 670bc325bd7bc3a1e12e6d07af3176ec
Run description: BM25, cohort constraint re-ranking, deep keyword deconstruction, no assertions; utilizes PubMed Central, Wikipedia, UMLS, and WordNet.

YorkUMB1¶

Run ID: YorkUMB1
Participant: york
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 30dbbafbabcf4787ab346e33a5e8798f
Run description: A baseline run retrieved by using BM25 algorithm with parameter b = 0.75 and k1 = 1.2. Top 2500 reports are selected and mapped to VisitID. The rank of a visit is determined by its top report. The returned visitIDs for each topic are less than 1000.

YorkUMC2¶

Run ID: YorkUMC2
Participant: york
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: 6a01664c8514458f55bbd0c7c61e271f
Run description: Automatic run. Use concept relationships in queries to improve the result of baseline run(BM25 b=0.75, k1=1.2). VisitID rank is determined by its top report in the retrieved report list.

YorkUMP4¶

Run ID: YorkUMP4
Participant: york
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: b8804edf5881400f9e5d4ce856a5f914
Run description: Using an extended Rocchio's feedback framework---the BM25 weighting model(b=0.3) + KL weighting for feedback (doc=10, term=30) + basic document weighting + full dependency proximity model (weight = 0.2)

YorkUMQ3¶

Run ID: YorkUMQ3
Participant: york
Track: Medical
Year: 2012
Submission: 8/7/2012
Type: automatic
Task: main
MD5: f65910ed3b22d2709b7640cabf74be2e
Run description: Using an extended Rocchio's feedback framework---the BM25 weighting model(b=0.3) + KL weighting for feedback (doc=10, term=30) + basic document weighting.