The majority of matched terms were single word terms and multi-word expressions.
In addition to these name decision rules, we also created several prefix and suffix rules to check whether a matched term is FP annotation.
Terms are tested for matches in descending order of length, so the longest possible matching term is always replaced with the appropriate CTERM or GTERM token.
There is no exact lowest level term for suicidal ideation in COSTART; the closest possible matching term is suicidal tendency, which is coded to the preferred term depression.
Matching report terms are replaced with their corresponding codes and terms from the nomenclature, and in a final step, all non-matching terms are replaced by a blocking tag.
Therefore, we can conclude that the dictionary-based pattern matching approach (Baseline II), i.e., tagging all terms by longest exact match in the test set if the matching terms were also tagged in the training set, is better than the TUI-based extraction approach (Baseline I).
In addition, the doublet autocoder successfully matched terms that were missed by the phrase autocoder, while the phrase autocoder found no terms that were missed by the doublet autocoder.
The chemical tokens and chemical name decision rules (introduced in Methods part 4) were used to decide if matched MeSH terms are full names or substrings.
Lexical parsers do not match terms that are absent from the text (i.e. no false positives).
The array of doublet runs extracted from the text that match nomenclature terms are cached in an external file.
