SENTENCE PATTERN CLASSIFICATION METHOD BASED ON MULTI COMBINATION KEYWORD OF SYLLABLES

摘要

In an automatic interpretation process for communication between people who use languages different from each other, the present invention relates to an automatic sentence pattern classification method based on a multiple combination keyword of syllables, in which voice recognition and automatic translation texts are analyzed, and a short sentence or a long sentence having complex meanings is automatically classified and a sentence pattern mark is added, based on multiple combination of all syllables configuring the text. The method includes the steps of: extracting a keyword candidate over a sentence pattern of each learning corpus, targeting a plurality of learning corpora; calculating a relative frequency rate of the extracted keyword candidate; comparing the calculated relative frequency rate to a preset critical value to select a keyword over the sentence pattern according to a comparison resu multiple-combining each syllable with respect to an input sentence according to a preset standard syllable number; and comparing the combined syllable and the selected keyword over the sentence pattern to classify the sentence pattern of a corresponding sentence, and to add a sentence pattern mark depending on the classified sentence pattern.