Table 1.
List of various features for the concept task
Category | Features |
Lexical context features | The target itself (n-gram) |
Syntactic context features | The POS |
The phrases of NPs and APs | |
Ontological features | UMLS-based dictionary matching |
MeSH-based dictionary matching | |
Medication dictionary matching | |
Head-noun dictionary matching | |
Sentence features | The sentence with a dosage-unit or temporal adverb (eg, twice daily or q4h) |
The sentence with numerals before dosage-units | |
The sentence containing a drug name before a numeral | |
The sentence with characteristics such as [the phrase] [numeral] [dosage] | |
The sentence with a drug name followed by an alternative drug name in parentheses, eg, fosamax (alendronate) | |
Word features | Word capitalized |
Entire word capitalized | |
Word abbreviation | |
The phrase before the numerals | |
Assertion word | |
Body word | |
Pattern matching | |
Word normalization | |
N-character-prefix-and-suffix | |
Word clustering |
AP, adjective phrase; NP, noun phrase; POS, part of speech.