摘要:本文运用规则和统计相结合的方法构造了一个汉语介词短语识别算法。首先,根据介词和介词短

语右边界组成的搭配模板自动提取可信搭配关系,并用这些搭配关系对介词短语进行识别。之后,用基于词

性的三元边界统计模型和规则相结合的方法识别其它未处理的介词短语。通过对含有7323 个介词短语的语

料作交叉测试,精确率达到87148 ％ ,召回率达到87127 ％。-Abstract: In this paper the use of rules and statistical methods of combining the structure of a prepositional phrase recognition algorithm for Chinese. First, according to preposition and prepositional phrase consisting of the right sector with the credibility of the template with automatic extraction of relations and relations with those with a prepositional phrase identification. After the POS-based statistical model of the ternary borders and rules of combination of the prepositional phrase to identify the other untreated. The adoption of 7323 prepositional phrase containing the corpus for cross-testing, accurate rate of 87,148 percent, the recall rate of 87,127 percent.