Chinese spelling checker in reading machine for visually impaired persons

The developing about supplementary instrument, the care giver for impaired persons, plays an essential role in orange technologies. This paper invested a Chinese spelling checker combining statistical and rule approaches for the reading machine. Word and part-of-speech levels n-gram models are adopted as the statistical module for detecting the spelling error and typos. Additionally, the rule set inducted by training corpus is aimed at capturing the patterns about spelling errors. According to the observations, the similar sets based on glyph and pronounce words are used as the features. By these two modules, the proposed approach can provide the spelling checking functionality in reading machine for visually impaired persons. As one of the major languages around the world, Chinese is becoming more and more important. However, the word-building and grammars are different compared to alphabetical languages. Therefore, Constructing of Chinese spelling checker is necessary and useful. According to experimental results, we can find that the proposed method can detection the spelling errors and provide the corresponding correction. The results show the proposed is practice and serviceable.