Word-level Analysis Task

A Word-level Analysis Task is a Linguistic Analysis Task that requires inferences about the Word Mentions in a Linguistic Expression.

• AKA: Word-level Analysis, Bag-of-Words Analysis.
• Context:
  • Input: A Linguistic Expression.
    • optional: A Target Word Mention.
  • output: Information about the Word Mentions (or Target Word Mention) in the Linguistic Expression.
  • It can be solved by a Word-level Analysis System (that implements a Word-level Analysis algorithm).
• Example(s):
  • a Word-level Syntactic Analysis Task.
    • a Part-of-Speech Tagging Task, requires a POS Label on each Word.
    • a Morphological Parsing Task, e.g. MPT("[Dogs] are mammals”) ⇒ (Stem dog] + Inflexional Suffix -s]
    • a Technical Term Mention Recognition Task.
  • a Word-level Semantic Analysis Task.
    • a Word Mention Normalization Task.
    • a Topic Modeling Task, requires the discovery of Topics in a Corpus, typically at the Word-level.
    • a Word-Sense Disambiguation Task, requires a WSD Label on each Word.
    • a Word Similarity Learning Task.
  • …
• Counter-Example(s):
  • a Morphological Rule Discovery Task.
  • a Relation Mention Detection Task.
  • a Sentence-level Analysis Task.
  • a Discourse-level Analysis.
  • a Syntactic Parsing Task.
  • a Semantic Parsing Task.
• See: Morpheme, Lexeme, Lexical Analysis, Annotated Document.

---

---

References

2002

• (Nair & Rost, 2002) ⇒ Rajesh Nair, Burkhard Rost. (2002). “Inferring Sub-cellular Localization Through Automated Lexical Analysis." Bioinformatics, 18(90001) doi:10.1093/bioinformatics/18.suppl_1.S78
  • QUOTE: We developed LOCkey, a fully automated method for lexical analysis of SWISS-PROT keywords that assigns sub-cellular localization.

---