Effective Use of Chinese Structural Auxiliaries for Chinese Parsing

Natural language parser is usually faced ungrammatical input, such as mistyping or error POS tags. If the parser uses language dependant explicit linguistic knowledge to detect and correct grammatical errors, it is useful for parser. In this paper, we propose a method that uses the Chinese structural auxiliary knowledge to detect and correct ungrammatical Chinese parsing errors. We focus on three error types: miss segmentation, miss POS tags, and miss typing. Experimental results show that appropriate use of evident Chinese structural auxiliary knowledge indeed helps to correct parsing errors and further to improve Chinese parsing performance.