面向社交网站的数据挖掘应用研究—用户关系分析

发布时间：2018-04-02 03:39

本文选题：社交网站　切入点：数据挖掘　出处：《南京邮电大学》2014年硕士论文

【摘要】：社交网站（Social Website，，简称SW）是现今互联网网站一个重要领域。社交网站提供一个网络平台，以供人们在社交网站上进行交友，通过与社交网站上的好友建立关系形成有关于用户的社交网络（Social Network Service），用户可以在社交网站的平台上与好友进行互动，增进之间的关系。随着移动互联网的发展，越来越多的人们能随时随地的在社交网站平台上发布、分享信息。社交网站正在扮演一个非常重要的消息传播平台。著名的社交网站有FaceBook、Twitter、YouTube、人人网、新浪微博等。由于在社交网站之中，用户是基于现实的好友建立起来的网络结构，因此它既有网络结构也具有社交的性质。本论文的研究重点分为三部分：第一部分分析社交网站中用户社交网络之间的用户关系。在对一般在线社交网站的网络属性和关系研究的基础上，针对社交网站的用户属性和行为进行分析。具体通过对微博网络中用户的粉丝数、关注数、评论信息数，转发信息数等研究，发觉社交网站用户关系网络的稀疏性、随机性、小世界性等重要性质。第二部分研究社交网站中用户在其社交网络中的影响力算法。根据前人对社交网络的用户影响力算法，深入研究PageRank算法思想。并从这个算法中分析出前人的算法仅仅是根据用户的出度和入度相关原理求得用户的影响力。于是，本文提出一种关于实体-关系的用户影响力算法，构建实体-关系影响力算法模型，提出影响用户影响力因素，通过PageRank算法思想进行计算。研究社交网站中用户的社交网络中，小型派系的检测。通过对已有的社区算法进行分析，在根据用户社交网络派系的特点，利用凝聚算法得到用户社交网络中的各个派系，以及各个派系中的中心人物。第三部分给出基于用户影响力算法和派系检测算法的基础上，介绍社交网站用户关系分析系统架构、数据库设计和系统界面设计。最终从社交网站中得到用户社交网络数据，得到用户社交网络中用户影响力排名前十的用户以及用户社交网络中的各个派系。
[Abstract]:Social networking site (SWE) is an important area of Internet today.Social networking sites provide an online platform for people to make friends on social networking sites.By building relationships with friends on social networking sites, there is a social network about the user, Social Network Service, which allows users to interact with friends on the platform of the social networking site to enhance the relationship between them.With the development of mobile Internet, more and more people can publish and share information on social networking platform anytime and anywhere.Social networking sites are serving as a very important messaging platform.Famous social networking sites are Facebook Twitter YouTube, Renren, Sina Weibo and so on.Because in the social network site, the user is based on the real friends to build the network structure, so it has both the network structure and the social nature.The research focus of this paper is divided into three parts:The first part analyzes the relationship between users' social networks in social networking sites.On the basis of the research on the network properties and relationships of the general online social networks, this paper analyzes the users' properties and behaviors of the social networks.By studying the number of users' fans, attention, comments and forwarding information in Weibo's network, the paper finds out the important properties of user relationship network of social networking sites, such as sparsity, randomness, small world and so on.In the second part, the influence algorithm of users in social network is studied.According to the previous user influence algorithm on social networks, the idea of PageRank algorithm is studied in depth.From this algorithm, we can find out that the former algorithm is only based on the principle of the correlation between the user's degree of output and the degree of entry to obtain the influence of the user.In this paper, a user influence algorithm about entity-relation is proposed, and the model of entity-relation influence algorithm is constructed, and the factors influencing user influence are put forward, and the calculation is carried out by the idea of PageRank algorithm.Study the detection of small factions in social networks by users of social networking sites.Based on the analysis of the existing community algorithms, according to the characteristics of the user's social network factions, the condensed algorithm is used to obtain each faction in the user's social network, as well as the central figure in each faction.The third part introduces the architecture, database design and interface design of user relationship analysis system based on user influence algorithm and factional detection algorithm.Finally, we get the user social network data from the social networking site, and get the top 10 users of the user social network and the various factions of the user social network.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP311.13;TP393.092

【相似文献】