As the era of social commerce is coming,the transactional community has become a key solution to the contradiction between economic nature and social nature in social commerce.Traditional social network analysis explored the evolution of social network,while transactional community is very different from social community regarding the roles and motivations of their members.This research,based on one of the most active communities in Taobao.com,analyzed the differences of transactional community in network closure mechanism.The results showed:(1) Members in transactional community would choose to avoid reciprocity because of high cost of social interaction and the risk of inefnciency of relationship;(2) In a vague situation where informational social influence impacts the most,more options of contagion path would enhance influence but harm necessity of its member,which has a negative effect on relationship estabUshment;(3) The relationship establishment in transactional community comes mainly from the similarities among members,which include the mutual acquaintances or mutual activities.

以淘宝网为平台,选取其中最活跃的圈子社区为研究对象,并将所有的社区关系区分为由买家发出的连接和由卖家发出的连接,分析交易型社区网络闭包机制相对于普通社交网络闭包机制的差异性以及社区中的买家和卖家在关系构建决策上的不同.研究结果表明:①交易型社区中的买家会回避互惠关系的形成,而卖家则会通过互惠性的连接来构建稳定的关系和培养忠诚顾客;②在信息性社会影响起主导作用的交易型社区中,买家用户会避免信息的重复和冗余性而追求信息的多样和丰富性,因此,作为信息源的用户的传染路径越多会降低该信息源的必要性从而不利于社区中的买家用户与其构建关系;另一方面,由于社区中的卖家用户希望通过具有较强传染性的用户来传播网络店铺或商品的相关信息,他们则更愿意与这些作为信息传染源的用户构建关系;③不论是对于买家用户还是卖家用户,交易型社区的网络闭包(关系构建)主要通过共同的社区好友和共同参与的社区活动这样的二模嵌入模式完成.

This study investigates the formation mechanism of dissemination force and the influence models of nodes’ network centrality in the virtual social network. Combining the Social Network Analysis and Tobit regression, we find that: 1) both a node’s degree centrality and betweenness centrality have a positive impact on its dissemination force; 2) the closeness centrality didn’t. For theory contribution, we have a clever understand of the source of member’ dissemination force and the various influence models of different node centralities in virtual social network. For practice contribution, different kinds of opinion leaders can be distinguished according to different centralities in a more accurate way, so that we can make a more effective use of their dissemination force in network marketing.

This study investigates the various influence models of nodes' network centrality in the context of transactional community. Combining the Social Network Analysis (SNA) with Tobit regression, the research indicates that: i) a node's degree centrality (its followers) and betweenness centrality (the number of the shortest paths in which the node is included) have a positive impact on its network influence, ii) the closeness centrality (physical closeness in network) shows no significant impact on its influence. Theoretically, the results provide insight into the sources of influence and various influence models of different node centralities in transactional community network. In practice, influentials can be better identified according to different centralities, so as to distinguish the opinion leaders in a more accurate way.