论文部分内容阅读
为精确描述问答网络论坛的主题信息,提出构建面向英语自然语言的问答知识表示模型,进而提出包括数据采集、平台搭建和应用设计在内的问答知识系统的完整解决方案.模型借助先验词典和自然语言语法分析方法,将语义信息、语法信息和知识置信度引入信息检索技术,并以语句模型集合和带有语义链接的标引项集合的形式表现出来.理论分析表明:语句模型能够突破原始问答模式限制,精确表达知识陈述,满足各种查询需求;标引项之间的语义链接则可辅助语句模型,在现有知识基础上推导出新知识.模型同时应用了信息检索和自然语言处理特征,增强了知识表达能力.诸多知识系统应用可在此模型上得到改善,提供更好的性能.
In order to accurately describe the theme information of Q & A online forum, a question and answer knowledge representation model oriented to English natural language is proposed, and then a complete solution to question and answer knowledge system including data collection, platform construction and application design is proposed. The model uses a priori dictionary and Natural language grammar analysis method, the semantic information, grammatical information and knowledge confidence into the information retrieval technology, and to the statement model set and semantic link with the set of indexing entries in the form of theoretical analysis shows that: the statement model can break through the original Querying and answering patterns to limit the precise expression of knowledge statements to meet the needs of a variety of queries; semantic links between the index items can be auxiliary statement model, based on the existing knowledge derived new knowledge model also applied information retrieval and natural language processing Features, enhance the ability of knowledge expression.Many knowledge system applications can be improved on this model to provide better performance.