This paper provides a method of the automatic annotation by means of the synthetic use of the grammatical information to realize thejudgment and automatic annotation of the semantic relationship of textual coherence. The distribution rules of the parts of speech are used toeliminate the non-conjunctions, and the potential conjunctions are tagged. The pattern of the textual coherence is obtained by the synthetic use of thecollocation distance, collocation strength and syntactic position after matching the pattern in the corpus of conjunctions. Based on the above data, thesemantic relationship is tagged. Experiment shows that the method is effective.%为实现篇章连贯语义关系的判定与自动标注,提出一种综合运用关联词多种语法信息的自动标注方法.该方法利用关联词的词性分布规则排除非关联词,标注出潜在关联词,对比关联词库中的模式表,并综合利用搭配距离、搭配强度和句法位置获取合法的篇章连贯模式,在此基础上标注出其语义关系.通过实验验证了该方法的有效性.
展开▼
