Sentence pattern classification method based on multi combination keyword of syllables

摘要

The present invention relates to an automatic interpretation process for communicating with people using different languages, in which speech recognition and automatic translation texts are analyzed to generate a long sentence having a short sentence or a multiple meaning based on multiple combinations of all syllables constituting the text The present invention relates to a syllable multiple combination keyword-based automatic classification method for automatically classifying syllables and adding a sentence pattern to a sentence classification keyword based on a syllable multiple combination keyword. The method includes extracting keywords candidates for each sentence type based on a plurality of learning corpuses; Calculating a relative frequency rate of each extracted keyword candidate; Comparing the calculated relative frequency rate with a predetermined threshold value, and selecting the keywords according to the resu Multiplexing each syllable with respect to a sentence input according to a predetermined number of reference syllables; And comparing the combined syllable with the selected keyword for each sentence type to classify the sentence sentence pattern and adding a sentence pattern according to the classified sentence pattern.