semantic similarity graph граф семантичної схожості когерентність тексту DOM Doc2Vec model HTML page template Trinity method UDC 004.67 UDC 004.83 automated collection of web pages classification of extraction methods coherence of a text convolutional neural network data extraction document discrimination task extraction methods formation of a regular expression insertion task natural language processing neural network prefix tree traversal recurrent neural network search bot semantic representation of sentences semantic similarity value of sentences ternary document tree text coherence УДК 004.67 УДК 004.83 автоматизированный сбор веб-страниц