sõltuvusanalüüsi

Sõltuvusanalüüsi, known in English as dependency analysis or dependency parsing, is a field within computational linguistics that focuses on determining the grammatical relationships between words in a sentence. Instead of analyzing sentences based on their linear order (like in constituency parsing), dependency analysis views sentences as a network of words where each word depends on another word, with the exception of the root word of the sentence. These relationships are typically represented as directed arcs connecting words, indicating which word modifies or governs another.

The core of dependency analysis involves identifying the head word and the dependent word for each relationship. For instance, in the sentence "The quick brown fox jumps over the lazy dog," "jumps" would be the main verb and likely the root of the sentence. "Fox" would depend on "jumps" as its subject, and "quick" and "brown" would depend on "fox" as its adjectives. The analysis aims to capture these structural dependencies, which can reveal deeper semantic and syntactic information about the sentence.

Different formalisms exist for dependency grammar, such as Universal Dependencies, which provides a framework for cross-linguistically consistent annotation of lexical and grammatical structure. The output of a dependency parser is often a dependency tree or a directed acyclic graph, illustrating these relationships. This analysis is crucial for various natural language processing tasks, including machine translation, information extraction, sentiment analysis, and question answering, as understanding the grammatical structure is fundamental to understanding the meaning of text.