Dependency Parsing

Dependency Parsing maps sentence tokens to labeled head-dependent pairs: it's the core syntactic analysis that determines who modifies or controls whom.

Dependency Parsing analyzes a sentence's grammatical structure, creating a directed, labeled dependency tree. Each word (token) is a node, and the edges are binary, asymmetric relations that link a 'head' word to its 'dependent.' For example, in 'The cat chased the mouse,' the parser identifies 'chased' as the root, 'cat' as its nominal subject (nsubj), and 'mouse' as its direct object (dobj). Modern systems, such as spaCy's transition-based parser, use neural models for high accuracy and speed, enabling critical downstream tasks like Information Extraction and Machine Translation.

https://spacy.io/usage/linguistic-features#dependency-parse