The retrieval research of non-adjacent keywords in Chinese corpus — A case study of “Yi…Jiu…” construction

Corpus Concordancing is a popular research topic. The function of retrieving data from corpus by providing non-adjacent keywords is widely used by users. However, the precision of retrieval results is not very high because the machine can´t recognize the relationship of the non-adjacent keywords. To deal with this problem, this paper proposed a rule-based method for the “Yi...Jiu...” construction, which could exclude the unrelated data, even though the data include the keywords. The experiments show that the precision is close to 82%.