冉晓斌,刘跃文,姜锦虎.社交网络个体活跃行为的大数据分析:从网络外部性的视角[J].管理科学,2017,30(5):77-86
社交网络个体活跃行为的大数据分析:从网络外部性的视角
Big Data Analysis of the Active Behavior in Social Networks:The perspective of Network Externality
投稿时间:2016-12-19  修订日期:2017-06-13
DOI:
中文关键词:  社交网络  持续使用  活跃行为  网络外部性  Tobit模型
英文关键词:social network  continue usage  active behavior  network externality  Tobit model
基金项目:国家自然科学基金(71301128,91546119,71331005)
作者单位E-mail
冉晓斌 西安交通大学 管理学院西安 710049 ranxiaobin@stu.xjtu.edu.cn 
刘跃文 西安交通大学 管理学院西安 710049 liuyuewen@mail.xjtu.edu.cn 
姜锦虎 西安交通大学 管理学院西安 710049 jiangjinhu@mail.xjtu.edu.cn 
摘要点击次数: 593
全文下载次数: 
中文摘要:
      用户是社交网络平台的基石,用户的活跃行为为平台创造内容、带来价值,因而用户活跃度一直是社交网络平台的关注重点。已有研究仅关注用户是否存在持续使用意向或行为,将持续使用行为定义为二元变量,未对持续使用行为细分,忽略了对用户活跃行为的研究。此外,社交网络中用户活跃行为之间有着不可忽视的相互影响,而受制于用户的关系网络数据难以获取,少有研究从该角度对社交网络中的用户行为进行研究。
将持续使用行为进一步细分为个体活跃和沉默行为,从网络外部性视角探讨同伴活跃行为对焦点个体活跃行为的影响以及个体网络规模与活跃行为的关系,并进一步探讨个体间关系强度和个体特征的调节作用。以超过百万用户的真实关系网络、个体特征信息和使用行为数据进行实证分析,通过Tobit模型验证上述关系。
研究结果表明,同伴活跃度和个体网络规模对个体活跃度有显著的正向作用,同时关系强度具有正向调节作用;个体特征在同伴活跃度与个体活跃度之间起调节作用,具体来说,女性用户和年轻用户受同伴活跃行为的影响更强。
从网络外部性的角度探讨用户活跃行为的影响因素,拓展了关于在线社交网络用户持续使用行为的研究和网络外部性测量方式,关注不同节点对于网络外部性贡献的差异性,对今后网络外部性的研究有一定的借鉴意义。基于中国数百万真实用户的数据集,结果更具普适性,同时反映出中国在线社交网络中用户的行为特征,可为今后的研究者提供参考。研究结果可以帮助社交网络平台的管理者更加了解用户活跃度的影响机理以及用户行为之间的内在联系,从而能以更科学合理的方式激发用户活跃度,保持社交网络平台的持续运营。
英文摘要:
      Users are the cornerstone of social network platforms. Users′ active behavior can generate content and add value for platforms. Therefore, user activation is always a great concern of social network platforms. Existing studies mainly focus on the intention or behavior of continuous usage of social network users. Moreover, they define continuous usage as the binary variable, which hinders the subdivision of continuous usage behavior and further ignores research on users′ active behavior. Although there is an undeniable interaction between users′ active behaviors in social network, few studies explored user behavior in social network from the perspective of peer effect due to the low accessibility of users′ social network data. Thus, the current study aims to fill this research gap.
Dividing continuous usage behavior into active and inactive behavior, this study explores the relationships between ①peers′ and egos′ active behavior; and ②users′ social network size and active behavior from the perspective of network externalities. Then, this study investigates the moderating effects of tie strength and individual characteristics. Finally, hypotheses are examined by Tobit model with a large scale social network dataset containing personal feature and individual behavior from more than one million users.
Results show that peers′ active behavior and individuals′ social network size both have a significant positive effect on individual activity, and tie strength has a positive moderating effect. In addition, individual characteristics have a moderating effect on the relationship between peers′ and users′ active behavior. Specifically, female and younger users are more sensitive to peers′ active behavior.
This study expands research on continuous usage behavior in social network by discussing factors that affect users′ active behavior from the perspective of network externality. It also extends the measurement of network externality. Different from prior research that uses network size(i.e., the number of nodes) as the network externality variable, this study focuses on the diverse contributions of various nodes to network externality, which may also inspire future research on network externality. Owing to the analysis of dataset containing millions of real users in China, the results become more universal and could reflect users′ behavior traits in Chinese online social network, which could provide references for future researchers. Meanwhile, these results can help managers of social network platforms understand the influence mechanism of user activity better and figure out the internal relationship between peers′ and users′ behaviors. Thus, they can stimulate user activity by a more scientific and reasonable way, and maintain the continuous operation of social network platforms.
查看全文  查看/发表评论  下载PDF阅读器
关闭