互联网用户偏好描述方法、形成机制与演化模式研究

项目名称： 互联网用户偏好描述方法、形成机制与演化模式研究

项目编号： No.61271308

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 无线电电子学、电信技术

项目作者： 沈波

作者单位： 北京交通大学

项目金额： 70万元

中文摘要： 随着互联网技术的日臻成熟和网络应用的日益丰富，越来越多的用户参与其中，也产生了大量具有应用价值的数据与信息。如何从大量繁杂的信息中快速而又准确地找到人们关心的内容，从数以亿计的用户中发现哪些用户可能对哪些信息感兴趣，是亟待解决和难以解决的问题，也是认识社会化网络演化规律和发掘互联网潜力与价值的基础，具有极大挑战性。实践证明，采用单纯的数据搜索和挖掘技术无法根本解决这些问题。鉴于此，本课题在前期研究的基础上，拟以交叉学科的方法和思想，采用信息科学的手段，从用户偏好的角度出发,通过对网络用户偏好的一般性描述方法、用户偏好模型与挖掘算法、偏好形成机制与演化模式、微观偏好与宏观偏好之间的联系以及大规模数据分析算法等问题的研究，揭示网络用户偏好及其演化的规律，从而建立有效的数据和用户过滤算法，为进一步挖掘互联网的应用潜力与价值，乃至推动网络舆论的研究和丰富复杂网络方面的理论，提供一些探索性结果。

中文关键词： 网络信息处理；信息提取；网络信息挖掘；检索模型；信息聚类

英文摘要： Plenty of mature network applications attract more and more users to the Internet and make the generating speed of network data to an inexpressible degree. And, in all these data there are great potential and value. How to fleetly and accurately find the content that people cares from a large amount of complicated information and dig out who might be interested in what kind of information from millions of network users. They are nasty and extreme challenging issues to be solved. They are also the foundation of understanding the law of social network evolution and mining the potential of Internet. However, it turns out that these issues cannot be solved by data searching and mining technologies purely. In this situation, the project will be focused on them from the angle of user preference and based on the preliminary study. In the project, the following issues will be studied by cross-disciplinary methods and thought, and particularly by the means of information science: general description approach of network user preference; model and mining algorithm of user preference; mechanism of preference formation and evolution mode; the relationship between microscopic preference and macroscopic preference; analysis algorithm of large scale data. Our purposes are to acknowledge and understand network user preference an

英文关键词： Internet information processing；information retrieval；Internet information mining；retrieval model；information clustering