融入紧密度中心性与信用的社交网络用户影响力强度计算模型  

A User Influence Strength Model in E-commerce Social Networks Based on Closeness and Users' Credit

作　　者：琚春华[1,2] 赵凯迪 鲍福光 Ju Chunhua;Zhao Kaidi;Bao Fuguang(Department of Modern Business Research Center, Zhejiang Gongshang University, Hangzhou 310018;School of Management Science & Engineering, Zhejiang Gongshang University, Hangzhou 310018)

机构地区：[1]浙江工商大学现代商贸研究中心,杭州310018 [2]浙江工商大学管理工程与电子商务学院,杭州310018

出　　处：《情报学报》2019年第2期170-177,共8页Journal of the China Society for Scientific and Technical Information

基　　金：浙江省哲学社会科学规划课题"基于消费者认知心理的行为决策与融合产品的社会化推荐应用研究"(16NDJC188YB);国家自然科学基金项目"电商环境下融入在线社会关系的消费信贷价值度量研究"(71571162);浙江省软科学重点项目"基于DEA的网上技术市场运行效率评价研究"(2018C25030)

摘　　要：在社交网络中意见领袖对信息传播有着巨大的促进作用,意见领袖往往能够影响群众以及引导网络舆论的走向。寻找网络中的意见领袖可以及时准确地掌握网络动态。本文提出一种融入紧密度中心性与信用的用户影响力强度计算模型,寻找电商化社交网络中的意见领袖。该模型首先根据用户间的好友关系获得关系邻接矩阵。然后用该邻接矩阵计算每个用户的紧密度中心性。提出SocialCreditRank算法计算用户影响力,该算法选择用户在网络中的紧密度中心性比重作为用户被随机选择的概率,用户与好友的信誉度之比对好友的贡献度进行修正。本文以支付宝的用户数据作为实验对象,实验结果表明该方法比一般的意见领袖识别方法效果更加精确。Opinion leaders play a major role in promoting information dissemination in social networks. Opinion leaders can often influence the masses and guide the trend of network public opinion. Looking for opinion leaders in the network can timely and can accurately grasp network dynamics. In this paper, we propose a calculation model of user influence intensity that integrates closeness centrality and tightness of credit, and looks for opinion leaders in e-commerce social networks. First, the model obtains the relationship adjacency matrix according to the friend relationship between users. Then the compactness centrality of each user is calculated with an adjacency matrix. The social credit rank algorithm is proposed for calculating user influence. The algorithm chooses the density centrality proportion of the user in the network as the probability that the user is randomly selected. The ratio of the user's credibility to that of the friend's contribution is revised. In this paper, the user data of Alipay is used as an experimental object. The experimental results show that the method is more accurate than the general opinion leader identification method.

关 键 词：电商化社交网络 意见领袖 影响力强度 紧密度中心性 

分 类 号：G206[文化科学—传播学]