Table Name |
Used for Scrubbing |
Used for Publication Analysis |
|---|---|---|
Feature_matrix_test |
YES |
|
Feature_matrix_train |
YES |
|
Human_annotations_test |
YES |
|
Human_annotations_train |
YES |
|
Machine_annotations_test |
YES |
|
Machine_annotations_train |
YES |
|
Lookup_dictionary |
YES |
|
Lookup_term_frequency |
YES |
|
Lookup_umls |
YES |
|
Pubs_authors |
|
YES |
Pubs_keywords |
|
YES |
Pubs |
|
YES |
Pubs_refs |
|
YES |
Feature_matrix_*
Stores feature matrix that is built from the Machine_annotations_* and Human_annotations_* tables. This is the rolled up feature set used for classification.
Human_annotations_*
Stores all annotations created by humans as part of a manual annotations effort.
Machine_annotations_*
Stores all annotations created by the UIMA pipeline.
Lookup_dictionary
Contains names from the 1990 US census that are used in
Lookup_term_frequency
Contains term frequency calculated across a random selection of 10,000 open access medical publications. Raw open access publications are available for free through NIH/NLM.
Lookup_umls
Contains terms from UMLS subset that was used Scrubber.
This DOES NOT include the UMLS CUIDs due to licensing restrictions.