基于社交网络模型的智能个人知识库构建

发布时间：2018-10-31 19:39

【摘要】：近些年来,人工智能技术快速发展,极大地促进了技术创新,以及推动了人类社会的发展。智能助手是随着人工智能的发展而兴起的一种新型的行动代理软件,受到了学术界和工业界广泛的关注。用户的兴趣具有动态性和易变性,为了提供智能化、个性化的服务,智能助手首要解决的两个问题是如何学习和适应用户兴趣。为了有效解决这两个问题,本文提出一种通过构建智能个人知识库来学习和适应用户兴趣的方案。本文的主要的研究内容包括以下几点:第一,深度调研智能助手的发展现状与研究热点,分析智能助手未来的发展趋势。深入了解开源智能助手项目Lucida及其工作原理。第二,分析维基百科的知识网络结构,并以其作为基础数据源,构建语义实体网络。以该语义实体网络作为通用知识库,提出基于社交网络模型构建智能个人知识库的方法。首先,提出面向用户兴趣的个人知识库的概念,并设计个人知识库的结构模型与构建流程;其次,描述用户兴趣量化建模方法;然后,提出基于语义实体网络的结构特征及文本特征的实体综合表征模型,并利用改进的基于内容的协同过滤算法和排序学习算法进行用户兴趣知识挖掘;最后,设计了个人知识库的自适应演化算法。第三,探索个人知识库在问答系统中的应用,分别从准确性和时效性两个方面,评测本课题构建的个人知识库原型。发现本课题所构建的个人知识库在一定程度上可以帮助理解用户的查询意图,提高问答的精确性;并且可以大大减少智能问答系统的响应时间,提高问答的时效性。验证了本课题提出的基于社交网络模型构建智能个人知识库方法的有效性。
[Abstract]:In recent years, the rapid development of artificial intelligence technology has greatly promoted technological innovation and the development of human society. Intelligent assistant is a new kind of mobile agent software, which is developed with the development of artificial intelligence. The interest of users is dynamic and changeable. In order to provide intelligent and personalized services, the first two problems solved by intelligent assistants are how to learn and adapt to the interests of users. In order to solve these two problems effectively, this paper proposes a scheme to learn and adapt to the interests of users by constructing an intelligent personal knowledge base. The main research contents of this paper are as follows: first, the development status and research focus of intelligent assistant in depth investigation, and analyze the development trend of intelligent assistant in the future. Learn more about open source intelligent helper project Lucida and how it works. Secondly, the knowledge network structure of Wikipedia is analyzed and used as the basic data source to construct semantic entity network. Taking the semantic entity network as a general knowledge base, a method of constructing intelligent personal knowledge base based on social network model is proposed. Firstly, the concept of personal knowledge base oriented to user interest is proposed, and the structure model and construction flow of personal knowledge base are designed, secondly, the quantitative modeling method of user interest is described. Then, an entity synthesis representation model based on semantic entity network and text feature is proposed, and the improved content-based collaborative filtering algorithm and ranking learning algorithm are used to mine user's interest knowledge. Finally, an adaptive evolutionary algorithm of personal knowledge base is designed. Thirdly, it explores the application of personal knowledge base in Q & A system, and evaluates the prototype of personal knowledge base from two aspects of accuracy and timeliness. It is found that the personal knowledge base constructed in this paper can help to understand the user's query intention to a certain extent, improve the accuracy of the question and answer, and greatly reduce the response time of the intelligent question answering system and improve the timeliness of the question and answer system. The validity of the intelligent personal knowledge base method based on social network model is verified.
【学位授予单位】：南昌大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18

【相似文献】