Chinese Spell Checking Based on Noisy Channel Model
Hsun-wen Chiu, Jian-Cheng Wu, Jason S. Chang · 2014
Chinese spell checking is an important component of many NLP applications, in-cluding word processors, search engines, and automatic essay rating. Compared to English, Chinese has no word bound-aries and there are various Chinese in-put methods that cause different kinds of typos, so it is more difficult to develop spell checkers for Chinese. In this paper, we introduce a novel method for correct-ing Chinese typographical errors based on sound or shape similarity. In our approach, similar characters are automatically gener-ated using Web corpora, and potential ty-pos in a given sentence are then corrected using a channel model and a character-based language model in the noisy channel model. In the training phase, we estimate the channel probabilities for each charac-ter based on ngrams in Web corpus. At run-time, the system generates correction candidates for each character in the given sentence and selects the appropriate cor-rection using the channel model and the language model. 1