商业视角下的网络社区的用户行为研究

发布时间：2018-03-26 01:08

本文选题：网络社区　切入点：用户行为　出处：《山东师范大学》2012年硕士论文

【摘要】：自上世纪九十年代起，网络社区随着因特网的飞速发展而产生，其发展速度之快引起很多学者的注意。但是学者们对于网络社区的定义始终没有达成一致，多数著作都认为网络社区是指包括BBS论坛、贴吧、公告栏、群组讨论、在线聊天、交友、个人空间、无线增值服务等形式在内的网上交流空间，同一主题的网络社区集中了具有共同兴趣的访问者。作为用户，我们只要工作学习中需要用到网络，就会不可避免的成为某个或某些个网络社区的成员。网络社区的独特优势及特性使得这个新生事物深入人心，网络社区的形成对于用户来说毫无疑问是有利的，不仅丰富了互联网生活，更为用户拓展了互联网空间，加深了共同兴趣方面的知识深度。网络社区的个数已经呈指数级增长，面对如此激烈的竞争，研究网络社区中的用户行为对于社区运营者来说是非常必要的，，首要原因就是商业利益。DCCI 2009-2010中国互联网市场数据显示，网络社区广告营收规模增速低于受众规模增长，2010年年底中国互联网社区论坛受众规模为1.83亿人，而到2011年年底这一数字大幅增加至2.45亿人，净增6200万人，增幅达33.9%。由此可见，谁能够争取到这部分用户，谁就能够获得更大的利益。因此，必须设置怎样的吸引量来引起用户的注意并留住用户使之产生的行为对网站最为有利。这就需要对用户的行为进行透彻的了解和彻底的分析，课题的研究具有实际意义和必要性。本文以QQ社区为研究对象，所设计问卷及调查分析均以QQ社区为模版。首先使用结构方程建模的方法对于影响网络社区盈利的因子进行归类，并对其产生的影响程度量化处理。结构方程建模是一种目前管理学研究中常用的数据分析方法。它是一种与多元回归分析关系密切，却在原理和方法上有许多拓展的多变量数据分析方法。它涵盖了多种原有的多变量数据分析方法，适用于定序、定类以及定距定比尺度，在管理、社会科学的实证研究中，逐渐成为与多元回归分析并立的一种主要多变量数据分析方法。根据用户对调查问卷的反馈，统计各选项出现频率，进而使用关联规则进行数据挖掘以达到对用户归类的目的，继而使用复杂网络中的社团划分算法找到社区中的中心用户和桥用户，对各类用户分别进行讨论，对于不同类型的用户要采取不同的维护策略，最终使QQ社区建设拥有更多更好的经济基础。复杂网络技术和数据挖掘技术都是比较成熟的技术，把这两种技术同时应用在网络社区的用户行为分析上是一种比较新的研究和应用。本文结合使用两种人工智能技术，为网络社区中用户行为的分析找到了一个新的角度，可以使网络社区网站清晰了解具有何种特征的用户会对网站产生正面影响以及影响的程度如何。
[Abstract]:Since the 1990s, with the rapid development of the Internet, the network community has attracted the attention of many scholars. However, scholars have never reached agreement on the definition of the network community. Most of the works think that the online community refers to online communication spaces, including BBS forums, posts, bulletin boards, group discussions, online chat, dating, personal space, wireless value-added services, and so on. The online community on the same topic brings together visitors of common interest. As users, we only need to use the network in our work and study. It will inevitably become a member of one or some network communities. The unique advantages and characteristics of the network community make this new thing deeply rooted in the hearts of the people. The formation of the network community is undoubtedly beneficial to the users. It not only enriches the Internet life, but also expands the Internet space and deepens the knowledge of common interest. The number of online communities has increased exponentially. In the face of such fierce competition, it is very necessary to study the behavior of users in the network community for community operators. The primary reason is that commercial interests. DCCI 2009-2010 China Internet market data show that the growth of advertising revenue in online communities is lower than the growth of audience size. At the end of 2010, the audience size of China Internet Community Forum was 183 million people. By the end of 2011, the number had increased significantly to 245 million, a net increase of 62 million, an increase of 33.9 percent. The behavior that must be set up to attract users' attention and retain them is most beneficial to the site. This requires a thorough understanding and thorough analysis of the user's behavior. The research of the subject has practical significance and necessity. This paper takes QQ community as the research object, designs the questionnaire and the investigation analysis takes the QQ community as the template. Firstly, classifies the factors that affect the profit of the network community by using the structural equation modeling method. Structural equation modeling is a data analysis method commonly used in management research at present. It is closely related to multivariate regression analysis. But there are many extended multivariate data analysis methods in principle and method. It covers many kinds of original multivariate data analysis methods, which are suitable for ordering, classifying, and distance determining scale, in the empirical research of management and social science. It has gradually become a main multivariate data analysis method parallel with multiple regression analysis. According to the feedback from the users to the questionnaire, the frequency of each option is counted, and then the association rules are used for data mining to achieve the purpose of classifying the users. Then the community partition algorithm in complex network is used to find the central users and bridge users in the community, and the different types of users are discussed, and different maintenance strategies are adopted for different types of users. Finally, QQ community construction has more and better economic base. The complex network technology and the data mining technology are both mature technologies. It is a relatively new research and application to apply these two technologies to the analysis of user behavior in the network community at the same time. A new angle is found for the analysis of user behavior in the network community, which can make the network community website clearly understand what kind of characteristics the users will have a positive impact on the website and the extent of the impact.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP393.09

【引证文献】