Corpus Refactoring: a Feasibility Study

Table 5 Roadmap for refactoring corpora. The list of corpora came from [32] and [33], where there are links to the corpora. Column headings indicate the steps that corpora may need to undergo to be refactored; those corpora that would require that step are noted with a dot. The heading "get original" means the original text needs to be retrieved. "Detect spans" means the corpus is a metadata corpus so spans of entities need to be detected. "Alt. search" means techniques other than exact-match searching must be used.

	get original	detect spans	alt. search
Arabidopsis Thaliana Circadian Rhythms [34]	•
Bio1 [35]	•
BioCreative 2004 Task 1A [28]	•		•
BioCreative 2004 Task 1B [36]		•	•
BioCreative 2004 Task 2 [37]		•	•
BioCreative 2006 Task GM [38]
BioCreative 2006 Task GN [39]
BioCreative 2006 Task IPS/IMS [40]		•	•
BioCreative 2006 Task ISS [40]		•
BioInfer [41]
BioText: Recognizing Abbreviation Defintions [42]
BioText: Protein-Protein Interaction Data [43]	•		•
BioText: Relations between Disease/Treatment Entities [44]	•
Brown-Genia Treebank [45]	•
DepGenia [46]	•
DIPPPI [47]		•	•
EDGAR [48]	•	•
GENIA [49, 50]	•
FetchProt [51]
Human Gene ID-Serve	•
IEPA [52]	•	•
ImmunoTome	•
iProLink [53]
Medstract [54, 55]
MedTag [7]
OHSUMED [56, 57]	•	•	•
PASBio [58]		•
PASTA [59]
PathBinder [60]
PennBioIE [12]
PICorpus
ProSpecTome [61]	•	•
PDG [9]	•	•	•
Texas [62]	•		•
TREC Genomics 2004 Categorization Task [63]		•	•
TREC Genomics 2005 Categorization Task [64]		•	•
TREC Gemonics 2006 IR Task [65]		•	•
TREC Genomics 2007 IR Task [65]		•	•
Wisconsin [66]	•	•	•
WSD [67]
Yapex [68, 69]	•