【作者】 王永芳；

【导师】 贾君枝；

【作者基本信息】 山西大学 ， 情报学， 2011， 硕士

【摘要】 当今网络上存在大量的与农民有关的信息资源,但大部分农民科技认知水平低,面对海量多样的信息资源,往往无从下手,无法找到自己需要的信息。此外,在当今流行的搜索引擎如Google、百度中能找到大量的与搜索项有关的信息,但其内容正确与否无法保证,指导农民实际应用存在一定风险。更重要的是,输入相关问题后这些搜索引擎返回的是相关的网页信息,而非简洁、准确的答案,这仍需要农民去阅读筛选,使用极其不方便。如何利用现有的农业信息及生活常识信息这些与农民密切相关的信息资源,针对农民的提问,提供准确、简洁的答案,使多数农民受益,是一个亟待解决的问题。面向农民的问答系统是将农业及生活常识类信息通过数据库形式组织起来,针对农民提出的问题,利用问答系统的思想提供准确的答案。本文的研究主要是基于贾君枝博士主持的山西省人文社科基地项目“山西省农民信息服务研究”(20093004)而展开。系统由四个模块组成：本地数据源构建、问句处理、信息检索和答案抽取。根据系统实现的步骤,本文分六章来研究。第一章分析国内外面向农民的问答系统的研究现状,提出本文研究的理论与现实意义。第二章阐述如何利用目前流行的网页抓取、分析及数据库构建技术构建检索的后台数据源。第三章分析收集的相关问句,总结特征,结合领域特点提出适合本系统的问句处理方式。第四章介绍了本系统信息检索的流程,提出了基于多特征加权融合的句子相似度计算方法。第五章根据问题的答案类型,将问题分为两大类(事实型问题和非事实型问题)进行答案抽取。第六章实现了一个面向农民的问答系统。本系统实现中使用了统计学相关方法,采用搜索引擎的关键技术和问答系统的核心技术。本文的研究结果是开发一个面向农民的问答系统,针对农民科技认知水平低的特点,为其提供简单、方便、快捷、有针对性的信息,使多数农民受益。本文研究具有一定的理论价值和实际应用价值。更多还原

【Abstract】 Nowadays, there are large amounts of information resources related to farmers on the network. However, most farmers’science and technology cognitive level is low, facing massive information resources, they are always unable to find information they need. Besides, although there are abundant information resources relevant to search items in modern popular search engines like Google and Baidu, it’s not necessarily correct to guide farmers’ practical application. More importantly, after searching, these search engines return relevant web pages rather than simple, accurate answers. Therefore, farmers still need to read and screen, which is extremely inconvenient. How to use existing information related to farmers such as agricultural information and information about daily life to provide accurate and simple answers, making most farmers benefited, is a problem to be solved. The farmer-oriented question answering system organizes agricultural information and information about daily life through database form, uses the thought of question answering system to provide accurate answers to farmers’problems.This research is mainly based on the Shanxi Humanities Social Sciences Base Project "The Study of Shanxi Farmers Information Service" (20093004), hosted by Doctor Jia Junzhi. The system of this study is composed of four modules:local data source construction, question processing, information searching and answer extraction. According to system implementation steps, this paper consists of six chapters. The first chapter analyzes the research status of farmer-oriented question answering system at home and abroad, and proposes theoretical and practical significance of this paper. The second chapter illustrates how to use currently popular technologies including web page grabbing, web page analysis and database construction to build retrieval data source. The third chapter analyzes some related questions, summarizes their features, and combines with field features; puts forward the question processing approach suited to this system. The forth chapter introduces the information retrieval process of this system, puts forward the calculation method for sentence similarity computing based on multi-features weighted fusion. The fifth chapter divides problems into two categories (factual questions and non-factual questions) for answering extraction, according to answer types of questions. The sixth chapter implements a farmer-oriented question answering system.During the realization of the system, we used some methods of statistics, key technologies of search engine and core technologies of question answering system. The aim of this paper is to develop a farmer-oriented question answering system, according to the farmers’low level of science and technology cognition, providing simple, convenient, quick, targeted information, benefiting most farmers. The research of this paper has certain theoretic value and practical application value.更多还原