基于依存关系的中文名词性谓词语义角色标注研究

摘要

语义角色标注是自然语言处理的一个重要研究内容,性能对机器翻译等研究有重大影响.实现了一个基于依存关系的中文名词性谓词语义角色标注平台,并对名词性谓词进行识别,使用最大熵分类模型在Chinese NomBank的转换语料上进行系统实验,对各种词法特征、结构特征及其组合进行了测试,标准语料上F1值达到78.09,基于自动句法树的语料上的F1值达到67.42.%Semantic Role Labeling is an important component of Natural Language Processing. It plays a critical role in machine translation. This paper explores a Chinese dependency-based Semantic Role Labeling system for Nominal Predicate by using a maximum entropy classifier. In particular, various kinds of lexical, syntactic and semantic features are incorporated to improve the performance with systematic evaluation on the dataset which is transferred from constituent-based corpus ( Chinese NomBank). Experiments with various syntactic or semantic feature and the combination of both prove the system can achieve 78.09 in labeled Fl for gold corpus and 67.42 for automatic parser tree.