The Research of Chinese Text Proofreading base on Pattern Matching

Chinese automatic proofreading of text is one of the main applications of natural language processing and a difficult problem for natural language understanding. Aiming at the errors caused by local or long-distance syntax or semantic constraints, this paper proposes a text proofreading method based on pattern matching. This method uses the characteristics of Chinese grammatical structure and semantic limitations to construct recognition pattern that integrates grammatical structure and conditional restrictions, and uses the defined pattern to match the sentences to be checked for error checking and error correction. Experiments show that the recall of the method reaches 95%, the accuracy reaches 90%, and the method has certain practicality.