NAME TRANSLITERATION METHOD BASED ON CLASSIFICATION OF NAME ORIGINS

摘要

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a translation system as a transliteration method based on human-origin classification. The present invention solves the problem of inconsistent transliteration modes in the Chinese-English name transliteration in different country origins. The method adopted in the present invention comprises: Classification of human origin, b. Linear interpolation system fusion. The method proposed in the present invention applies a logistic multi-classification regression model to the classification of human origin, classifies human origin based on the characteristic template of the human character constituent character, And the system fusion is performed on the results of several transliteration models to realize the bilingual translation.