Paragraph-oriented methods for determining the coherence and thematic unity of scientific and technical texts

The subject of the article is to determine degree scientific and technical text connectedness using statistical calculations. aim investigation study possibilities coherence fluctuations in relative frequencies keywords paragraphs lexical thematic unity texts. task develop a method for determining at set level; test developed methods on collection documents. used are analysis computational experiment methods. following results were obtained. has shown that it advisable cluster as points keyword space level paragraphs. This opens up possibility calculating within clusters entire text. fragments whole determined by analyzing sequence paragraph numbers clusters. makes possible formally quality material presented or textbook. Conclusions. novelty follows: there was refined determination (coherence unity) texts implementation clustering space, calculation inside overall text, well through order coherence. language-independent, based clear hypotheses, complement each other. have an adjusting element can be adapt different stylistic areas. It been experimentally proved proposed efficient provide framework information technology content do not use WEB resources syntactic semantic analysis, providing them autonomously.