Concept - Parsing
Definition (s.53): the step of producing a full parse of a sentence; each word is connected to a single structure, usually a tree. Parsers are built from statistical analysis of hand-parsed tree banks.
Why use an expensive process (s.53): it yields detailed syntactic relationship information that phrase identification cannot.
Phrase structure tree (s.53-54): S -> NP VP, etc. For “Johnson was replaced at XYZ Corp. by Smith” the tree lets a machine infer Johnson was replaced by Smith (correct), not Steve was replaced by Smith, in “Johnson who is son of Steve was replaced at XYZ Corp. by Smith” (s.54).
Dependencies (s.55-56): see Concept - Universal Dependencies. “I like him”: nsubj(like, I), dobj(like, him).
Tools: Stanford Parser (statistical parser, Java; online demo), Stanford CoreNLP (s.46, s.48, s.55-56); off-the-shelf tools (NLTK, CoreNLP, Spacy, Textblob) used in ~80% of applications (s.57).
Related
Concept - POS Tagging | Concept - Penn Treebank | Concept - Universal Dependencies | Concept - Sentence Boundary Detection