基于基字识别的藏文音节字检错算法研究

在藏文音节字构件中基字占有重要的地位，只要准确的识别出基字的位置，可通过音节字的长度来获得其他构件，并以基字为核心来判断组成音节宇其他构件的搭配是否符合规则。本文通过研究藏文音节字构件组合语法，提出了一种基于基字识别的藏文音节字检错算法。该方法不使用藏文音节字词典及大规模语料库的支撑。通过实验，该方法自动检错藏文音节字的正确率为97％。

Research on an detection algorithm of error Tibetan syllable based on Jizi recognition

Jizi plays an important role in the Tibetan syllable component. So long as the position of Jizi is recognized correctly, we can get other components through the length of syllable and take Jizi as core to judge whether the collocation of other components, which forms syllahie, conforms to rules can detcet error Tibetan syllable.This article proposes an error detection algorithm of Tibetan syllable based on lizi recognition by researching the combination grammar of Tibetan syllable component. And this method doesn＇t get any support of Tibetan syllable dictionaries and large-scale corpus. Through the experiments, this method＇s accuracy to detect Tibetan syllable automatically is 97% .