APPARATUS FOR GENERATING HIGH FREQUENCY VOCABULARY STRING RECOGNITION UNIT IN A DIALOGUE AND READ LARGE VOCABULARY SPEECH RECOGNITION SYSTEM AND METHOD THEREFOR

摘要

The present invention utilizes a high frequency pseudomorphic sequence as one recognition unit, and generates a high frequency lexical sequence recognition unit of a conversational and reading object large vocabulary continuous speech recognition system for generating a recognition unit of an intermediate form between a pseudomorpheme and a word. And to a method thereof. As described above, the present invention provides a frequency information extraction unit 301 for extracting continuous lexical pair frequency information from a pseudo morpheme-tagged text corpus, and the frequency information extracted by the frequency information extraction unit 301 and the length of each lexical pair. After modifying the text corpus based on the combined vocabulary selection unit 302 for selecting a vocabulary to be combined based on the information, and the vocabulary set selected by the combined vocabulary selection unit 302, a high frequency continuous vocabulary pair To generate a high frequency lexical sequence recognition unit based on the pseudo morpheme combining information corrector 303 that combines the unicode into one to generate a modified text corpus, and the text corpus generated by the pseudo morpheme combined information corrector 303. And a unit generating unit 304.;Conversations and readings Large vocabulary, text corpus, vocabulary dictionary, language model, pronunciation dictionary