基于隐式协同的社会化搜索排序研究

发布时间：2018-06-29 05:50

本文选题：社会化搜索 + SimRank　；参考：《哈尔滨工程大学》2013年硕士论文

【摘要】：互联网经过几十年的发展，已经极大程度上融入到了人们的现实生活当中，随着产业与需求的发展，互联网被划分为几大入口，包括搜索引擎、浏览器、即时通信以及当下流行的社会网络等等。搜索引擎解决了人们在互联网海量信息当中快速便捷地获取有效内容的问题，社会网络在虚拟网络世界建立了类现实的人际关系网络，拉近了人与人之间的距离。搜索引擎与社会网络作为两大互联网入口，不能孤立发展。传统搜索引擎对任何用户的相同搜索请求都会返回相同搜索结果，在进行个性化服务转型过程，搜索引擎往往只是根据用户兴趣等因素对用户单独的个性化服务，用户彼此的个性化信息不能够被相互借鉴。社会网络为用户相互借鉴个性化信息提供了良好的基础平台，用户在进行搜索时不再是孤军奋战，，而协同好友共同完成一次搜索任务。搜索引擎与社会网络的融合，催生了社会化搜索的相关研究。然而，社会化搜索的研究还处于一个起步阶段，研究都对于社会化搜索如何将搜索引擎与社会网络结合起来都有不同的认识。本文从社会网络可为搜索引擎提供协同式服务的角度出发，基于隐式协同对社会化搜索排序进行深入研究。本文的主要研究工作包括以下几个方面： 1.采用社会网络分析法对搜索引擎进行日志分析，以不确定图的方式逻辑表示搜索引擎的日志中查询词和网页的链接关系，通过基于不确定图的SimRank算法，计算查询词与网页的相似度，最终以相似度和查询词的加权方式建立网页描述库。 2.从分析用户搜索经验入手，计算社会网络中用户的信任度。在建立用户间信任度量的基础上提出隐式协同模型。 3.结合前两方面工作，综合提出社会化搜索排序算法。
[Abstract]:After decades of development, the Internet has been greatly integrated into people's real life. With the development of industry and demand, the Internet has been divided into several portals, including search engines, browsers, Instant messaging and the current popularity of social networks and so on. Search engine solves the problem that people can get effective content quickly and conveniently in the mass information of Internet. Social network has set up a kind of realistic interpersonal network in the virtual network world, which brings people closer to each other. Search engine and social network as two big Internet entrance, cannot develop in isolation. The traditional search engine will return the same search result to any user with the same search request. In the process of personalized service transformation, the search engine usually only individualizes the user according to the user's interest and other factors. Users' personalized information cannot be used for reference. Social network provides a good basic platform for users to learn from personalized information, users are no longer alone in searching, and cooperate with friends to complete a search task. The fusion of search engine and social network has given birth to the relevant research of social search. However, the research of social search is still in its infancy, which has different understanding on how to combine social search engine with social network. From the point of view that social network can provide collaborative service for search engine, this paper makes a deep research on social search ranking based on implicit collaboration. The main research work of this paper includes the following aspects: 1. The social network analysis method is used to analyze the search engine log. The query words in the search engine log and the link relationship between the web page and the query word in the search engine log are logically represented by the uncertain graph, and the SimRank algorithm based on the uncertain graph is adopted. The similarity between query words and web pages is calculated. Finally, the web page description library is established by similarity and weight of query words. 2. Based on the analysis of user search experience, the trust degree of users in social network is calculated. Based on the establishment of trust between users, an implicit cooperative model is proposed. 3. Combined with the first two aspects of work, a comprehensive social search sorting algorithm is proposed.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.3

【参考文献】