Automatic Text Analysis

(AA), the operation of extracting, from a given text in a natural language, the grammatical and semantic information contained in the text; the operation follows some algorithm in conformity with a description—elaborated in advance—of the particular language. The reverse operation is called automatic text synthesis. Automatic text analysis proceeds in three phases: (1) lexical-morphological—the transition from an individual word form to its lexical and grammatical characteristics; (2) syntactic—the transition to the syntactic structure of a sentence from the chain of lexical and grammatical characteristics representing it; (3) semantic—the transition from the syntactically analyzed sentence to the recording of its meaning. The algorithm of automatic text analysis is usually divided into information about the language (grammar) and information about the process of analysis itself (the “mechanism,” or algorithm proper of the analysis). Automatic text analysis is a necessary phase in different types of automatic processing of texts: automatic translation, automatic reference work, information search, and the like. Automatic text analysis should be distinguished from automatic text research; in the latter information about the language of the text is absent or almost absent, and the text is processed through algorithms with a view to developing a description of the language.