【摘要】 在线词汇资源FrameNet的语义类型对于深层次的自然语言理解具有重要的作用。文章首先对FrameNet的语义类型功能进行论述,指出其由框架类型、本体类型和词汇类型构成,从概念、感情、搭配等层面描述词汇固有的词类特征和语义特征。在此基础上,作者探讨语义类型在自然语言处理中的具体应用范围:如限制框架元素的取值范围,进行问答、文摘、复述和翻译等自然语言理解,便于概念检索中关系相关度的计算,实现语义层面的推理,建立与其他本体在多领域中的映射等。更多还原

【Abstract】 As an online lexical resource,the semantic types of FrameNet play an important role in deep natural language understanding. This paper first discusses the functions of the semantic types of FrameNet,pointing out that it consists of frame types,ontological types and lexical types. Semantic types describe the inherient lexical or semantic features of a word in perspectives of concept,sentiment and collocation. Based on this analysis,the author explores the concrete applications of semantic types in natural language precessing such as limiting the range of the fillers of frame elements in a frame and performing natural language understanding tasks such as question and answer,abstract,paraphrase and translation,thereby facilitating the calculation of ralationship similarity in concept retrieval to realize semantic reasoning and develop mutiple domain-specific links with other ontologies.更多还原