Runs - Genomics 2005¶

ABBR003¶

Run ID: ABBR003
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Allele task, using bayesian regression with full text representation and MH_Mesh check.

ABBR003SThr¶

Run ID: ABBR003SThr
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Allele task, First, use bayesian regression with full text representation and MH_Mesh check. Second, use output from SVM for sanity check.

ABPLUS¶

Run ID: ABPLUS
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Abstract plus mesh headings plus chemicals

ABPLUSE¶

Run ID: ABPLUSE
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Abstracts plus mesh headings plus chemicals E

ABPLUSG¶

Run ID: ABPLUSG
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Abstracts plus mesh headings plus chemicals G

ABPLUST¶

Run ID: ABPLUST
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Abstracts plus mesh headings plus chemicals T

aDIMACSg9md¶

Run ID: aDIMACSg9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Article title, abstract and MEDLINE MeSH terms were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained and used to make predictions with maximum expected effectiveness threshold tuning.

aDIMACSg9w¶

Run ID: aDIMACSg9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

aDIMACSl9md¶

Run ID: aDIMACSl9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

aDIMACSl9w¶

Run ID: aDIMACSl9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

aDUTCat1¶

Run ID: aDUTCat1
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Titles, absracts, bodies are used to represent the documents, backspace-separated words as features, a svm classifier and set weights using tfidf. All texts are processed with A Biomedical Named Entity Recognizer(abner).

aDUTCat2¶

Run ID: aDUTCat2
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Titles, absracts, bodies are used to represent the documents, backspace-separated words as features, a svm classifier and set weights using tfidf.

aFduMarsI¶

Run ID: aFduMarsI
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM

aFduMarsII¶

Run ID: aFduMarsII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM, specified feature

aFduMarsIII¶

Run ID: aFduMarsIII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, MeSH Tree Extracted Knowledge, Rocchio

Afull¶

Run ID: Afull
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: bag-of-words maximum entropy classifier

aibmadz05m1¶

Run ID: aibmadz05m1
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with probability thresholding.

aibmadz05m2¶

Run ID: aibmadz05m2
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with cross-validation thresholding.

aibmadz05s¶

Run ID: aibmadz05s
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Single classifiers.

aIBMIRLmet¶

Run ID: aIBMIRLmet
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using SVM the document set obtained as positive after the rules run (vectors mesh terms + class specific terms using the dic file).

aIBMIRLrul¶

Run ID: aIBMIRLrul
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: applying rules (mesh term and class specific term in medline record), formed using the training set, on the document set obtained after MH - Mice screening

aIBMIRLsvm¶

Run ID: aIBMIRLsvm
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using SVM of documents (vectors mesh terms + class specific terms using the dic file) obtained after MH - Mice screening

aLRIk1¶

Results | Participants | Input | Summary | Appendix

Run ID: aLRIk1
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic

aLRIk2¶

Results | Participants | Input | Summary | Appendix

Run ID: aLRIk2
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + list.

aLRIk3¶

Results | Participants | Input | Summary | Appendix

Run ID: aLRIk3
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + rap + list.

Ameta¶

Run ID: Ameta
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: for each document, the posterior probabilites of individiual paragraphs -- as assigned by a primary maximum entropy classifier -- are tallied as a distribution, which is then classified by a secondary maxent classifier

aMUSCUIUC1¶

Run ID: aMUSCUIUC1
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageA task with basic SVM classification

aMUSCUIUC2¶

Run ID: aMUSCUIUC2
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageA task with SVM classification with semantic features

aMUSCUIUC3¶

Run ID: aMUSCUIUC3
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageA task with SVM classification with semantic features and augmented training cases.

aNLMB¶

Run ID: aNLMB
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Mixture

aNLMF¶

Run ID: aNLMF
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Voting

aNTUMAC¶

Run ID: aNTUMAC
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: Mesh30+Caption10+Abstract10+CaptionSEM10+AbstractSEM10

AOHSUBF¶

Run ID: AOHSUBF
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Best-single feature classifier

AOHSUSL¶

Run ID: AOHSUSL
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Slipper2 classifier

AOHSUVP¶

Run ID: AOHSUVP
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Voting perceptron classifier

Apars¶

Run ID: Apars
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: averaging the posterior probabilities of selected individual paragraphs as assigned by a maximum entropy classifier trained on full documents

aQUNB8¶

Run ID: aQUNB8
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the feature selection was done on each resulting thematic cluster. The selected features were used for the classification into "in"/"out" with respect to Allele category.

aQUT11¶

Run ID: aQUT11
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the resulting thematic clusters were then used for classification into "in"/"out" with respect to Allele category.

aQUT14¶

Run ID: aQUT14
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the resulting thematic clusters were then used for classification into "in"/"out" with respect to Allele category.

asubaral¶

Run ID: asubaral
Participant: arizonau.baral
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: manual
Task: adhoc
Run description: Briefly list the most salient features of this run.

ASVMN03¶

Run ID: ASVMN03
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Allele task, use support vector machine with full text representation (tuned threshold) and MH_Mesh check.

aUCHSCnb1En3¶

Run ID: aUCHSCnb1En3
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. More strict feature selection.

aUCHSCnb1En4¶

Run ID: aUCHSCnb1En4
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. Less strict feature selection.

aUCHSCsvm¶

Run ID: aUCHSCsvm
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with Support Vector Machine.

CCP0¶

Run ID: CCP0
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: manual
Task: adhoc
Run description: LocusLink gene synonym expansion; stemming; topic-specific keyword expansion; UMLS for disease synonyms, with heavy manual filtering of synonyms for "cancer"; weighted title over abstract.

CCP1¶

Run ID: CCP1
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: manual
Task: adhoc
Run description: LocusLink gene synonym expansion, converting all synonyms to a "bag of words", weighting individual words by frequency in synset; stemming; topic-specific keyword expansion; UMLS for disease synonyms, also converted to BOW; weighted title over abstract.

cuhkrun1¶

Run ID: cuhkrun1
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic classification using context feature engineering and context association.

cuhkrun1E¶

Run ID: cuhkrun1E
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic classification using context feature engineering and context association.

cuhkrun1G¶

Run ID: cuhkrun1G
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic classification using context feature engineering and context association.

cuhkrun1T¶

Run ID: cuhkrun1T
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic classification using context feature engineering and context association.

cuhkrun2A¶

Run ID: cuhkrun2A
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances.

cuhkrun2E¶

Run ID: cuhkrun2E
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances.

cuhkrun2G¶

Run ID: cuhkrun2G
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances.

cuhkrun2T¶

Run ID: cuhkrun2T
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances.

cuhkrun3A¶

Run ID: cuhkrun3A
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances with all features.

cuhkrun3E¶

Run ID: cuhkrun3E
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances with all features.

cuhkrun3G¶

Run ID: cuhkrun3G
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances with all features.

cuhkrun3T¶

Run ID: cuhkrun3T
Participant: cuhk.lam
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Customized learning by exploiting the symmetric property of attributes and instances with all features.

dcu1¶

Run ID: dcu1
Participant: dublincityu.gurrin
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: This run is the result of pseudo-relevance feedback on a baseline obtained with the DCU CDVP search engine Fisreal. Our search engine implements the BM25 probabilistic algorithm and the pseudo-relevance feedback is using Robertson Offer Weight method. The feedback aims at expanding the original queries with terms related to the generic structure of the queries, i.e. the Generic Topic Templates. The expansion terms were extracted from the sample search results using the relevance judgment provided.

dcu2¶

Run ID: dcu2
Participant: dublincityu.gurrin
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: This run is the result of pseudo-relevance feedback on a baseline obtained with the DCU CDVP search engine Fisreal. Our search engine implements the BM25 probabilistic algorithm and the pseudo-relevance feedback is using Robertson Offer Weight method. The feedback aims at expanding the original queries with terms related to the generic structure of the queries, i.e. the Generic Topic Templates (GTTs). The top 5 documents for each topic of the same GTT are assumed relevant and GTT-related structure terms are extracted from these documents to expand topics that are instances of that particular GTT.

dpsearch1¶

Run ID: dpsearch1
Participant: datapark.zakharov
Track: Genomics
Year: 2005
Submission: 7/27/2005
Type: manual
Task: adhoc
Run description: The DataparkSearch engine of upcoming 4.32 version has been used with fast method and default options of relevancy calculation.

dpsearch2¶

Run ID: dpsearch2
Participant: datapark.zakharov
Track: Genomics
Year: 2005
Submission: 7/27/2005
Type: manual
Task: adhoc
Run description: The DataparkSearch engine of upcoming 4.32 version has been used with full method and default options of relevancy calculation.

DUTAdHoc1¶

Run ID: DUTAdHoc1
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 7/30/2005
Type: manual
Task: adhoc
Run description: Our ad hoc task retreival system mainly includes the following features gene synonym expansion,medical term expansion based on the Metathesaurus of UMLS Knowledge Sources provided by NLM,different scoring strategy on different parts of Medline record(Title,Abstract,RN,MH,etc.).

DUTAdHoc2¶

Run ID: DUTAdHoc2
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 7/30/2005
Type: manual
Task: adhoc
Run description: Our ad hoc task retreival system mainly includes the following features gene synonym expansion,medical term expansion based on the Metathesaurus of UMLS Knowledge Sources provided by NLM,different scoring strategy on different parts of Medline record(Title,Abstract,RN,MH,etc.)and Pseudo-relevant feedback.

EBBR0006¶

Run ID: EBBR0006
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Expression task, use bayesian regression with full text representation (tuned threshold) and MH_Mesh check.

EBBR0006SThr¶

Run ID: EBBR0006SThr
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Expression task, First use bayesian regression with full text representation (tuned threshold) and MH_Mesh check. Second, use output from SVM for sanity check.

eDIMACSg9md¶

Run ID: eDIMACSg9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Article title, abstract and MEDLINE MeSH terms were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained and used to make predictions with maximum expected effectiveness threshold tuning.

eDIMACSg9w¶

Run ID: eDIMACSg9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

eDIMACSl9md¶

Run ID: eDIMACSl9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Article title, abstract and MEDLINE MeSH terms were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained and used to make predictions with maximum expected effectiveness threshold tuning.

eDIMACSl9w¶

Run ID: eDIMACSl9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

eDUTCat1¶

Run ID: eDUTCat1
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Titles, absracts, bodies are used to represent the documents, only nouns as features, a svm classifier and set weights using tfidf. All texts are processed with A Biomedical Named Entity Recognizer(abner).

eDUTCat2¶

Run ID: eDUTCat2
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Titles, absracts, bodies are used to represent the documents, backspace-separated words as features, a svm classifier and set weights using tfidf.

eFduMarsI¶

Run ID: eFduMarsI
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM

eFduMarsII¶

Run ID: eFduMarsII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM, specified feature

eFduMarsIII¶

Run ID: eFduMarsIII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, MeSH Tree Extracted Knowledge, Rocchio

Efull¶

Run ID: Efull
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: bag-of-words maximum entropy classifier

eibmadz05m1¶

Run ID: eibmadz05m1
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with probability thresholding.

eibmadz05m2¶

Run ID: eibmadz05m2
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with cross-validation thresholding.

eibmadz05s¶

Run ID: eibmadz05s
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Single classifiers.

eIBMIRLmet¶

Run ID: eIBMIRLmet
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using svmlight the document set obtained as positive after the rules run (vectors mesh terms + class specific terms using the dic file).

eIBMIRLrul¶

Run ID: eIBMIRLrul
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: applying rules (mesh term and class specific term in medline record), formed using the training set, on the document set obtained after MH - Mice screening

eIBMIRLsvm¶

Run ID: eIBMIRLsvm
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using SVM of documents (vectors mesh terms + class specific terms using the dic file) obtained after MH - Mice screening

eLRIk1¶

Results | Participants | Input | Summary | Appendix

Run ID: eLRIk1
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: automatic

eLRIk2¶

Results | Participants | Input | Summary | Appendix

Run ID: eLRIk2
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + list.

eLRIk3¶

Results | Participants | Input | Summary | Appendix

Run ID: eLRIk3
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + Rap + list.

Emeta¶

Run ID: Emeta
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: for each document, the posterior probabilites of individiual paragraphs -- as assigned by a primary maximum entropy classifier -- are tallied as a distribution, which is then classified by a secondary maxent classifier

eMUSCUIUC1¶

Run ID: eMUSCUIUC1
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageE task with SVM classification with semantic features and augmented training cases. Records without MeSH term mice are removed

eMUSCUIUC2¶

Run ID: eMUSCUIUC2
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageE task with SVM classification with semantic features.

eMUSCUIUC3¶

Run ID: eMUSCUIUC3
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageE task with SVM classification with semantic features and augmented training cases. Record without MeSH mice are also included.

eNLMF¶

Run ID: eNLMF
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Voting

eNLMKNN¶

Run ID: eNLMKNN
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Mixture

eNTUMAC¶

Run ID: eNTUMAC
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: Mesh10+Caption10+CaptionSEM10

EOHSUBF¶

Run ID: EOHSUBF
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Best-single feature classifier

EOHSUSL¶

Run ID: EOHSUSL
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Slipper2 classifier

EOHSUVP¶

Run ID: EOHSUVP
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Voting perceptron classifier

Epars¶

Run ID: Epars
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: averaging the posterior probabilities of selected individual paragraphs as assigned by a maximum entropy classifier trained on full documents

eQUNB11¶

Run ID: eQUNB11
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the feature selection was done on each resulting thematic cluster. The selected features were then used for classification into "in"/"out" with respect to Expression category.

eQUNB19¶

Run ID: eQUNB19
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The naive Bayes approach for classification into "in"/"out" with respect to Expression category. The different costs on different misclassifications were taken into account.

eQUT18¶

Run ID: eQUT18
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the resulting thematic clusters were then used for classification into "in"/"out" with respect to Expression category

ESVMN075¶

Run ID: ESVMN075
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Expression task, use support vector machine with full text representation (tuned threshold) and MH_Mesh check.

eUCHSCnb1En3¶

Run ID: eUCHSCnb1En3
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. More strict feature selection.

eUCHSCnb1En4¶

Run ID: eUCHSCnb1En4
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. Less strict feature selection.

eUCHSCsvm¶

Run ID: eUCHSCsvm
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with Support Vector Machine.

FTA¶

Run ID: FTA
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Fulltext A

FTE¶

Run ID: FTE
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Fulltext E

FTG¶

Run ID: FTG
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Fulltext G

FTT¶

Run ID: FTT
Participant: erasmus.kors
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: Fulltext T

GAbsBBR0083¶

Run ID: GAbsBBR0083
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Go annotation task, use bayesian regression with abstract text representation (derived threshold) and MH_Mesh check.

GBBR004¶

Run ID: GBBR004
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Go annotation task, use bayesian regression with full text representation (tuned threshold) and MH_Mesh check.

gDIMACSg9md¶

Run ID: gDIMACSg9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Article title, abstract and MEDLINE MeSH terms were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained and used to make predictions with maximum expected effectiveness threshold tuning.

gDIMACSg9w¶

Run ID: gDIMACSg9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

gDIMACSl9md¶

Run ID: gDIMACSl9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Article title, abstract and MEDLINE MeSH terms were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained and used to make predictions with maximum expected effectiveness threshold tuning.

gDIMACSl9w¶

Run ID: gDIMACSl9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

gDUTCat1¶

Run ID: gDUTCat1
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Documents are classified indepedently by three parts, its PMID record's MeSH term, full text, and title plus abstract. Then the results are combined by a evaluation program. We use a svm classifier, backspace-separated words as features, a tfidf method to set weights. Full texts are processed by A Biomedical Named Entity Recognizer(abner). Finally the positive instances are filtered with a algorithm according to the numbers of proteins of each document.

gDUTCat2¶

Run ID: gDUTCat2
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Documents are classified indepedently by three parts, its PMID record's MeSH term, full text, and title plus abstract. Then the results are combined by a evaluation program. We use a svm classifier, backspace-separated words as features, a tfidf method to set weights. Full texts are processed by A Biomedical Named Entity Recognizer(abner).

genome1¶

Run ID: genome1
Participant: csusm.guillen
Track: Genomics
Year: 2005
Submission: 7/31/2005
Type: automatic
Task: adhoc
Run description: We used the INDRI system developed by UMASS and CMU to create five indexes. Then we used the "runquery" option to retrieve documents using the five indexes. The topics were mapped to the INDRI format before retrieving the documents.

genome2¶

Run ID: genome2
Participant: csusm.guillen
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: We used the INDRI system developed by UMASS and CMU to create five indexes. Then we used the "runquery" option including the feedback parameter with 100 documents to retrieve documents. The topics were mapped to the INDRI system format before running the queries.

gFduMarsI¶

Run ID: gFduMarsI
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM

gFduMarsII¶

Run ID: gFduMarsII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM, specified feature

gFduMarsIII¶

Run ID: gFduMarsIII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, MeSH Tree Extracted Knowledge, Rocchio

Gfull¶

Run ID: Gfull
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: bag-of-words maximum entropy classifier

gibmadz05m1¶

Run ID: gibmadz05m1
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with probability thresholding.

gibmadz05m2¶

Run ID: gibmadz05m2
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with cross-validation thresholding.

gibmadz05s¶

Run ID: gibmadz05s
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Single classifiers.

gIBMIRLmet¶

Run ID: gIBMIRLmet
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using SVM the document set obtained as positive after the rules run (vectors mesh terms + class specific terms using the dic file).

gIBMIRLrul¶

Run ID: gIBMIRLrul
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: applying rules (mesh term and class specific term in medline record), formed using the training set, on the document set obtained after MH - Mice screening

gIBMIRLsvm¶

Run ID: gIBMIRLsvm
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using SVM of documents (vectors mesh terms + class specific terms using the dic file) obtained after MH - Mice screening

gLRIk1¶

Results | Participants | Input | Summary | Appendix

Run ID: gLRIk1
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic

gLRIk2¶

Results | Participants | Input | Summary | Appendix

Run ID: gLRIk2
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + list.

gLRIk3¶

Results | Participants | Input | Summary | Appendix

Run ID: gLRIk3
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + List + Rap .

Gmeta¶

Run ID: Gmeta
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: for each document, the posterior probabilites of individiual paragraphs -- as assigned by a primary maximum entropy classifier -- are tallied as a distribution, which is then classified by a secondary maxent classifier

gMUSCUIUC1¶

Run ID: gMUSCUIUC1
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageG task with SVM classification with full vocabulary.

gMUSCUIUC2¶

Run ID: gMUSCUIUC2
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageG task with SVM classification with sematnic feature. Parameter 'j' is set to 8.

gMUSCUIUC3¶

Run ID: gMUSCUIUC3
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageG task with SVM classification with sematnic feature. Parameter 'j' is set to 20.

gNLMF¶

Run ID: gNLMF
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Voting

gNTUMAC¶

Run ID: gNTUMAC
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: Mesh10+Caption10+MeshSEM10

GOHSUBF¶

Run ID: GOHSUBF
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Best-single feature classifier

GOHSUSL¶

Run ID: GOHSUSL
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Slipper2 classifier

GOHSUVP¶

Run ID: GOHSUVP
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Voting perceptron classifier

Gpars¶

Run ID: Gpars
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: averaging the posterior probabilities of selected individual paragraphs as assigned by a maximum entropy classifier trained on full documents

gQUNB12¶

Run ID: gQUNB12
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The naive Bayes approach was used for classification into "in"/"out" with respect to GO category. The different costs on different misclassifications were taken into account.

gQUNB15¶

Run ID: gQUNB15
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the feature selection was done on each resulting thematic cluster. The selected features were then used for classification into "in"/"out" with respect to GO category.

gQUT22¶

Run ID: gQUT22
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the resulting thematic clusters were then used for classification into "in"/"out" with respect to GO category

GSVMN08¶

Run ID: GSVMN08
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Go annotation task, use support vector machine with full text representation (tuned threshold) and MH_Mesh check.

gUCHSCnb1En3¶

Run ID: gUCHSCnb1En3
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. More strict feature selection.

gUCHSCnb1En4¶

Run ID: gUCHSCnb1En4
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. Less strict feature selection.

gUCHSCsvm¶

Run ID: gUCHSCsvm
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with Support Vector Machine.

i2r1¶

Run ID: i2r1
Participant: iir.yu
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Automatic run of Institute for Infocomm Research.

i2r2¶

Run ID: i2r2
Participant: iir.yu
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: The 2nd run from Institute for Infocomm Research

iasl1¶

Run ID: iasl1
Participant: academia.sinica.tsai
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Without using query expansion in Template 1. Without relevance feedback.

iasl2¶

Run ID: iasl2
Participant: academia.sinica.tsai
Track: Genomics
Year: 2005
Submission: 8/2/2005
Type: automatic
Task: adhoc
Run description: With query expansion in Template 1 and relevance feedback

ibmadz05bs¶

Run ID: ibmadz05bs
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Primary run. Enhanced automatic relevance feedback. Synonyms from external resources. Queries enhanced by bi-grams.

ibmadz05us¶

Run ID: ibmadz05us
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Secondary run. Enhanced automatic relevance feedback. Synonyms from external resources.

iitprf011003¶

Run ID: iitprf011003
Participant: iit.urbain
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Modivied pvn with 1 iteration RF.

LPC6¶

Results | Participants | Input | Summary | Appendix

Run ID: LPC6
Participant: langpower.yang
Track: Genomics
Year: 2005
Submission: 8/27/2005
Task: categorization
Run description: System built using rule-based NLP parsing and concept-based indexing.

LPC7¶

Results | Participants | Input | Summary | Appendix

Run ID: LPC7
Participant: langpower.yang
Track: Genomics
Year: 2005
Submission: 8/27/2005
Task: categorization
Run description: System built using rule-based NLP parsing and concept-based indexing.

MARYGEN1¶

Run ID: MARYGEN1
Participant: umaryland.oard
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: InQuery with proximity operators for phrases identified using MetaMap and disease name expansion using MetaMap

NCBIMAN¶

Results | Participants | Input | Summary | Appendix

Run ID: NCBIMAN
Participant: nlm.wilbur
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: manual
Task: adhoc
Run description: Same as NCBITHQ for all queries except template 1 (100-109) which used manual theme generation.

NCBITHQ¶

Results | Participants | Input | Summary | Appendix

Run ID: NCBITHQ
Participant: nlm.wilbur
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Phrases and their variants are extracted from each query and used to form boolean queries. The non-gene resuls are expanded using a "theme" approach (naive Bayes scoring) to rescore the results. Document scores from individual queries are converted to probabilites and combined with fuzzy logic operations. Template 1 uses the MEDLINE nearest neighbor function instead of boolean queries on separate phrases, and combines the results with a generic "protocol theme".

NLM1G¶

Run ID: NLM1G
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Mixture

NLM1T¶

Run ID: NLM1T
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Mixture

NLM2A¶

Run ID: NLM2A
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: ML methods

NLM2E¶

Run ID: NLM2E
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: ML methods

NLM2G¶

Run ID: NLM2G
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: ML methods

NLM2T¶

Run ID: NLM2T
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: ML methods

NLMfusionA¶

Run ID: NLMfusionA
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: combination of four systems NCBI, Smart, InQuery, EZIR with query expansion

NLMfusionB¶

Run ID: NLMfusionB
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: combination by template of four systems NCBI, Smart, InQuery, EZIR with query expansion

NTUgah1¶

Run ID: NTUgah1
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 7/30/2005
Type: automatic
Task: adhoc
Run description: The Entrez Gene and MeSH databases are used to identify important topic terms and their synonymns. For a topic, documents are first ranked by whether they contain all the important terms, and than by BM25 scores.

NTUgah2¶

Run ID: NTUgah2
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 7/30/2005
Type: automatic
Task: adhoc
Run description: Same as NTUgah1, except that documents which contain all important terms in their abstracts or titles are ranked higher than those which contain important terms appearing only in their MeSH fields.

OHSUall¶

Run ID: OHSUall
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 7/29/2005
Type: automatic
Task: adhoc
Run description: Briefly list the most salient features of this run.

OHSUkey¶

Run ID: OHSUkey
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 7/29/2005
Type: automatic
Task: adhoc
Run description: Briefly list the most salient features of this run.

PDnoSE¶

Results | Participants | Input | Summary | Appendix

Run ID: PDnoSE
Participant: upadova.bacchin
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: This is a TF-IDF vector based IR system.

PDSESe02¶

Results | Participants | Input | Summary | Appendix

Run ID: PDSESe02
Participant: upadova.bacchin
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: The IR system uses a query expansion technique based on symbol recognition.

SFUshi¶

Run ID: SFUshi
Participant: simon-fraseru.shi
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: manual
Task: adhoc
Run description: 1. Make use of public gene/protein database to expand query; 2. Use structured query to express logic relations among query terms; 3. Use pseudo relevance feedback;

TBBR0004¶

Run ID: TBBR0004
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Tumor task, use bayesian regression with full text representation (derived threshold) and MH_Mesh check.

TBBR0004SThr¶

Run ID: TBBR0004SThr
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Tumor task, First use bayesian regression with full text representation (derived threshold) and MH_Mesh check. Second, use output from SVM for sanity check.

Tcsusm1¶

Run ID: Tcsusm1
Participant: csusm.guillen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The documents went through two passes. In the first pass we selected those documents containing the keywords as they appear in Area tumor column "Always select" (Library Triage Cheat Sheet). In a second pass we discarded those documents including Cell Lines for mouse and human (Library Triage Cheat Sheet). The decision making process in both passes is done with decision rules

Tcsusm2¶

Run ID: Tcsusm2
Participant: csusm.guillen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: A third pass was added at the end for this run. In this pass we filtered the documents left from the second pass using the keywords not to choose documents. The decision making process was also done with decision rules.

tDIMACSg9md¶

Run ID: tDIMACSg9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Article title, abstract and MEDLINE MeSH terms were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained and used to make predictions with maximum expected effectiveness threshold tuning.

tDIMACSg9w¶

Run ID: tDIMACSg9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Gaussian prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

tDIMACSl9md¶

Run ID: tDIMACSl9md
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Article title, abstract and MEDLINE MeSH terms were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained and used to make predictions with maximum expected effectiveness threshold tuning.

tDIMACSl9w¶

Run ID: tDIMACSl9w
Participant: rutgers.dayanik
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Fulltext obtained from journal articles' subject, title, abstract and body fields were used for document representation. Text representation stemmed, logtf-idf, cosine normalization. A classifier, Bayesian Binary Regression (BBR) with Laplace prior, was trained, and used to make predictions with maximum expected effectiveness threshold tuning.

tDUTCat1¶

Run ID: tDUTCat1
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Titles, absracts, bodies are used to represent the documents, backspace-separated words as features, a svm classifier , set weights using tfidf, all texts are processed with A Biomedical Named Entity Recognizer(abner).

tDUTCat2¶

Run ID: tDUTCat2
Participant: dalianu.yang
Track: Genomics
Year: 2005
Submission: 8/19/2005
Task: categorization
Run description: Titles, absracts are used to represent the documents, backspace-separated words as features, a svm classifier , set weights using tfidf.

tFduMarsI¶

Run ID: tFduMarsI
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM

tFduMarsII¶

Run ID: tFduMarsII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, Certain Domain Retrieval Otherness Model, SVM, specified feature

tFduMarsIII¶

Run ID: tFduMarsIII
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Porter Stemmer, Pubmed Stopword, MeSH Tree Extracted Knowledge, Rocchio

Tfull¶

Run ID: Tfull
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: bag-of-words maximum entropy classifier

THUIRgA0p9x¶

Run ID: THUIRgA0p9x
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: CV version.

THUIRgen1S¶

Run ID: THUIRgen1S
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Structural Query Language; UniSentence, BiSentence and Multi-field Retrieval; Internal Resource Utility; Iterative Result Fusion; Stemming, Stopword, BM2500, Pseudo-relevance feedback; (More details will be involved in our report.)

THUIRgen2P¶

Run ID: THUIRgen2P
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Pattern Generation; Pattern Matching and Scoring; Prefix, Midfix, and Suffix for Given Template Expansion; Balance Between Precision and Recall; Internal Resource Utility. (More details will be involved in our report.)

THUIRgenA1p1¶

Run ID: THUIRgenA1p1
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 8/31/2005
Task: categorization
Run description: Baseline run. Only distanced bigram features.

THUIRgenE1p8¶

Run ID: THUIRgenE1p8
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 8/31/2005
Task: categorization
Run description: Baseline run. Only distanced bigram features.

THUIRgenG1p1¶

Run ID: THUIRgenG1p1
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 8/31/2005
Task: categorization
Run description: Baseline run. Only distanced bigram features.

THUIRgenGMNG¶

Run ID: THUIRgenGMNG
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: n-gram features. MeSH lib added.

THUIRgenT1p5¶

Run ID: THUIRgenT1p5
Participant: tsinghua.ma
Track: Genomics
Year: 2005
Submission: 8/31/2005
Task: categorization
Run description: Baseline run. Only distanced bigram features.

tibmadz05m1¶

Run ID: tibmadz05m1
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with probability thresholding.

tibmadz05m2¶

Run ID: tibmadz05m2
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Combined classifiers with cross-validation thresholding.

tibmadz05s¶

Run ID: tibmadz05s
Participant: ibm.zhang
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Single classifiers.

tIBMIRLmet¶

Run ID: tIBMIRLmet
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using svmlight the document set obtained as positive after the rules run (vectors mesh terms + class specific terms using the dic file).

tIBMIRLrul¶

Run ID: tIBMIRLrul
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: applying rules (mesh term and class specific term in medline record), formed using the training set, on the document set obtained after MH - Mice screening

tIBMIRLsvm¶

Run ID: tIBMIRLsvm
Participant: ibm-india.ramakrishnan
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: classifying using SVM documents (vectors mesh terms + class specific terms using the dic file) obtained after MH - Mice screening

tLRIk1¶

Results | Participants | Input | Summary | Appendix

Run ID: tLRIk1
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic

tLRIk2¶

Results | Participants | Input | Summary | Appendix

Run ID: tLRIk2
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + list.

tLRIk3¶

Results | Participants | Input | Summary | Appendix

Run ID: tLRIk3
Participant: uparis-sud.kodratoff
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Automatic + list + rap.

Tmeta¶

Run ID: Tmeta
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/2/2005
Task: categorization
Run description: for each document, the posterior probabilites of individiual paragraphs -- as assigned by a primary maximum entropy classifier -- are tallied as a distribution, which is then classified by a secondary maxent classifier

tMUSCUIUC1¶

Run ID: tMUSCUIUC1
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageG task with SVM classification with sematnic feature. Parameter 'j' is set to 20.

tMUSCUIUC2¶

Run ID: tMUSCUIUC2
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageG task with SVM classification with sematnic feature and augmented training cases. Parameter 'j' is set to 20.

tMUSCUIUC3¶

Run ID: tMUSCUIUC3
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/30/2005
Task: categorization
Run description: This is triageG task with SVM classification with sematnic feature and augmented training cases. Parameter 'j' is set to 20. Records without MeSH terms under the subtree 'neoplasma'.

tNLMF¶

Run ID: tNLMF
Participant: nlm-umd.aronson
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Voting

tnog10¶

Results | Participants | Input | Summary | Appendix

Run ID: tnog10
Participant: tno.erasmus.kraaij
Track: Genomics
Year: 2005
Submission: 7/31/2005
Type: automatic
Task: adhoc
Run description: JM smoothed language model

tnog10p¶

Results | Participants | Input | Summary | Appendix

Run ID: tnog10p
Participant: tno.erasmus.kraaij
Track: Genomics
Year: 2005
Submission: 7/31/2005
Type: automatic
Task: adhoc
Run description: JM smoothed language model, + Journal title prior

tNTUMAC¶

Run ID: tNTUMAC
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: Mesh30+Caption30+Abstract30+AbstractSEM10+CaptionSEM30

tNTUMACasem¶

Run ID: tNTUMACasem
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: AbstractSEM10

tNTUMACwj¶

Run ID: tNTUMACwj
Participant: ntu.chen
Track: Genomics
Year: 2005
Submission: 8/24/2005
Task: categorization
Run description: Mesh30+Caption30+wjAbstract30+AbstractSEM10+CaptionSEM30

TOHSUBF¶

Run ID: TOHSUBF
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Best-single feature classifier

TOHSUSL¶

Run ID: TOHSUSL
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Slipper2 classifier

TOHSUVP¶

Run ID: TOHSUVP
Participant: ohsu.hersh
Track: Genomics
Year: 2005
Submission: 8/22/2005
Task: categorization
Run description: abstract title, word, MeSH, and gene tokens Chi-square feature selection Voting perceptron classifier

Tpars¶

Run ID: Tpars
Participant: uwisconsin.craven
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: averaging the posterior probabilities of selected individual paragraphs as assigned by a maximum entropy classifier trained on full documents

tQUNB3¶

Run ID: tQUNB3
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The naive Bayes approach was used for for classification into "in"/"out" with respect to Tumor category. The different costs on different misclassifications were taken into account.

tQUT10¶

Run ID: tQUT10
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the resulting thematic clusters were then used for classification into "in"/"out" with respect to the Tumor category.

tQUT14¶

Run ID: tQUT14
Participant: queensu.shatkay
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: The probabilistic approach was used for creating probabilistic models of thematic clusters of all the irrelevant documents first, and the resulting thematic clusters were then used for classification into "in"/"out" with respect to the Tumor category.

TSVM0035¶

Run ID: TSVM0035
Participant: ibm.kanungo
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Tumor task, use support vector machine with full text representation (tuned threshold) and MH_Mesh check.

tUCHSCnb1En3¶

Run ID: tUCHSCnb1En3
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. More strict feature selection.

tUCHSCnb1En4¶

Run ID: tUCHSCnb1En4
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with NaiveBayes. Less strict feature selection.

tUCHSCsvm¶

Run ID: tUCHSCsvm
Participant: ucolorado.cohen
Track: Genomics
Year: 2005
Submission: 9/1/2005
Task: categorization
Run description: Categorized based on select bigrams, extracted strain names, and mesh terms. Classifications performed with Support Vector Machine.

UAmscombGeFb¶

Run ID: UAmscombGeFb
Participant: uamsterdam.aidteam
Track: Genomics
Year: 2005
Submission: 7/29/2005
Type: automatic
Task: adhoc
Run description: Run that combines MeSH-heading based blind feedback with gene name synonym and acronym expansion

UAmscombGeMl¶

Run ID: UAmscombGeMl
Participant: uamsterdam.aidteam
Track: Genomics
Year: 2005
Submission: 7/29/2005
Type: automatic
Task: adhoc
Run description: Run that combines gene name synonym and acronym expansion with automatic MeSH-heading lookup procedure

UBIgeneA¶

Run ID: UBIgeneA
Participant: suny-buffalo.ruiz
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Automatic run using gene expansion with MeSH terms, minimal stemming, and restricted word bigrams. IR system SMART, weighting scheme atn.ann

UICgen1¶

Run ID: UICgen1
Participant: uillinois-chicago.liu
Track: Genomics
Year: 2005
Submission: 7/29/2005
Type: automatic
Task: adhoc
Run description: Porter stemming; OKAPI; Query expansion; Weighting scheme

UIowa05GN101¶

Run ID: UIowa05GN101
Participant: uiowa.eichmann
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Precision focussed run. Uses a more stringent threshold on the ranked results.

UIowa05GN102¶

Run ID: UIowa05GN102
Participant: uiowa.eichmann
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Recall focussed run.

UIUCgAuto¶

Run ID: UIUCgAuto
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/2/2005
Type: automatic
Task: adhoc
Run description: This run is produced completely automatically from the original topic description. It performs pseudo feedback based on the structure of the query using a language modeling approach.

UIUCgInt¶

Run ID: UIUCgInt
Participant: uiuc.zhai
Track: Genomics
Year: 2005
Submission: 8/2/2005
Type: interactive
Task: adhoc
Run description: This run is produced with human relevance judgments on the top 20 documents from the initial retrieval run. It also uses biology resources to automatically expand the original queries.

UMD01¶

Run ID: UMD01
Participant: umichigan-dearborn.murphey
Track: Genomics
Year: 2005
Submission: 7/27/2005
Type: automatic
Task: adhoc
Run description: We extracted key words from each topic and combine keywords with logic connection AND or OR. We then calculated similarity scores of all documents with this combination and sorted the results.

UMD02¶

Run ID: UMD02
Participant: umichigan-dearborn.murphey
Track: Genomics
Year: 2005
Submission: 7/30/2005
Type: automatic
Task: adhoc
Run description: We extracted key words from each topic,then calculated similarity scores by Okapi BM25 method and sorted the results.

UniGe2¶

Run ID: UniGe2
Participant: u.geneva
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: This runs merges two different lists 1) a run with query expansion, based on gene and protein names and rocchio; 2) a run with expansion based on MeSH terms. Warning this run is intended to replace the run 'UniGeC' ! Please, run 'UniGeC' is corrupted and should be deleted.

UniGeNe¶

Run ID: UniGeNe
Participant: u.geneva
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: Data fusion (combination of two result lists) by a) a probabilistic model + pseudo-relance feedback (10 docs / 20 terms) b) same probabilistic model with modified queries (with thesaurus of gene and protein names) + PRF (10 docs / 20 terms)

UniNeHug2¶

Run ID: UniNeHug2
Participant: uneuchatel.savoy
Track: Genomics
Year: 2005
Submission: 7/27/2005
Type: automatic
Task: adhoc
Run description: Probabilistic model + pseudo-relance feedback (10docs / 20 terms)

UniNeHug2c¶

Run ID: UniNeHug2c
Participant: uneuchatel.savoy
Track: Genomics
Year: 2005
Submission: 7/27/2005
Type: automatic
Task: adhoc
Run description: Data fusion (combination of two result lists) by a) a probabilistic model + pseudo-relance feedback (10docs / 20 terms) b) same probabilistic model with modified queries (with genomics DB) + PRF (10 docs / 20 terms)

uta05a¶

Run ID: uta05a
Participant: utampere.pirkola
Track: Genomics
Year: 2005
Submission: 7/17/2005
Type: automatic
Task: adhoc
Run description: This is a simple run that serves as a baseline for our second run. Topic keys were used in queries, no expansion etc. was used. Different columns of a template were linked by a Boolean conjunction.

uta05i¶

Run ID: uta05i
Participant: utampere.pirkola
Track: Genomics
Year: 2005
Submission: 7/29/2005
Type: interactive
Task: adhoc
Run description: 1. Synonymous gene names for the topic gene names were retrieved from the Entrez Gene. 2. Our automatic queries (uta05a) were expanded with the synonyms. 3. The expanded queries were run on the test database. 4. Final queries (Boolean queries) were formulated by further expanding the queries with MH terms and synonyms found in the top documents of the initial search.

uwmtEg05¶

Results | Participants | Input | Summary | Appendix

Run ID: uwmtEg05
Participant: uwaterloo.clarke
Track: Genomics
Year: 2005
Submission: 7/4/2005
Type: automatic
Task: adhoc
Run description: Plain Okapi BM25 run, with stemming applied to all terms that do not contain numerical characters.

uwmtEg05fb¶

Results | Participants | Input | Summary | Appendix

Run ID: uwmtEg05fb
Participant: uwaterloo.clarke
Track: Genomics
Year: 2005
Submission: 7/4/2005
Type: automatic
Task: adhoc
Run description: Okapi BM25 run with standard Okapi feedback; stemming applied to all terms that do not contain numerical characters. This is a two-stage run, using the top 40 documents returned by the first stage to add pseudo-relevance feedback terms to the query in the second stage.

wim1¶

Run ID: wim1
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: language model,greek letter,query expansion

wim2¶

Run ID: wim2
Participant: fudan.niu
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: gene noun,okapi,query expansion

YAMAHASHI1¶

Results | Participants | Input | Summary | Appendix

Run ID: YAMAHASHI1
Participant: utokyo.takahashi
Track: Genomics
Year: 2005
Submission: 8/2/2005
Type: manual
Task: adhoc
Run description: Using MeSH for ranking.

YAMAHASHI2¶

Results | Participants | Input | Summary | Appendix

Run ID: YAMAHASHI2
Participant: utokyo.takahashi
Track: Genomics
Year: 2005
Submission: 8/2/2005
Type: manual
Task: adhoc
Run description: Not Using MeSH for ranking.

york05ga1¶

Run ID: york05ga1
Participant: yorku.huang
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: 1. use Okapi BM25 System with stuctured query function 2. use rules to expand the terms. 3. use BioNLP utility to identify the long form and acronym pairs. 4. use some rules to rebalance the weight for query term. 5. blank feedback with special term selection technique

york05ga2¶

Run ID: york05ga2
Participant: yorku.huang
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: 1. Use Okapi BM25 2. Blank feedback with term selection technique.

york05ga3¶

Run ID: york05ga3
Participant: yorku.huang
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: automatic
Task: adhoc
Run description: voting from the results from different systems. different sources are given the equal weight to vote. this run is voted by york05ga1 and york05ga2.

york05ga4¶

Run ID: york05ga4
Participant: yorku.huang
Track: Genomics
Year: 2005
Submission: 8/2/2005
Type: automatic
Task: adhoc
Run description: 1. use Okapi BM25 System with structured query function 2. use Acromed and LocusLink database to expand the terms 3. automatically delete redundant terms. 4. rebalance the weight for terms in the query 5. blank feedback with special term selection technique.

york05ga5¶

Run ID: york05ga5
Participant: yorku.huang
Track: Genomics
Year: 2005
Submission: 8/2/2005
Type: automatic
Task: adhoc
Run description: 1. use Okapi BM25 System with structured query function 2. use Acromed and LocusLink database to expand the terms, for some term not in database, fuzzy search has been used 3. automatically delete redundant terms. 4. rebalance the weight for terms in the query 5. blank feedback with special term selection technique.

york05gm1¶

Run ID: york05gm1
Participant: yorku.huang
Track: Genomics
Year: 2005
Submission: 8/1/2005
Type: manual
Task: adhoc
Run description: 1. use Okapi BM25 system with structured query function 2. use Acromed and LocusLink database to expand the terms. 3. mannualy select good quality expanded terms 4. use some rules to rebalance the weight for query term 5. blank feedback with special term selection technique