Metadata augmented via NLP tools such as PubTator are bound to have context-dependent erroneous entries. For example, PERCH in one record may be some sort of technique that Pubtator will extract as a fish, and in another record PubTator will correctly extract perch as a fish.
The augmented metadata is currently stored in a SQLite database as a table relating terms and ontology ids. A table with record ids will be needed
Metadata augmented via NLP tools such as PubTator are bound to have context-dependent erroneous entries. For example, PERCH in one record may be some sort of technique that Pubtator will extract as a fish, and in another record PubTator will correctly extract perch as a fish.
The augmented metadata is currently stored in a SQLite database as a table relating terms and ontology ids. A table with record ids will be needed