《计算机应用研究》|Application Research of Computers

在线用户打分行为长记忆效应与信任关系研究

Study of long-term memory in online rating behavior for trust formation

免费全文下载 (已被下载 次)  
获取PDF全文
作者 郭昕宇,郭强,刘建国
机构 上海理工大学 复杂系统科学研究中心,上海 200093
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2019)08-007-2275-04
DOI 10.19734/j.issn.1001-3695.2018.03.0094
摘要 对在线打分行为的动态研究能够帮助深入理解社交网络用户集群行为和信任关系的演化机制,当前许多在线系统用户能够通过对物品进行打分传达自己的观点。通过去趋势波动分析法研究了用户打分行为在信任关系建立前后的长记忆效应,并通过随机化打分时间和信任时间建立零模型,最后进行用户打分行为异质性分析。采用Epinions数据集进行实证研究,结果表明用户打分的长记忆效应在信任关系建立前出现下降趋势(8.06%),并于之后逐步回升(8.43%),而在两个零模型中赫斯特指数分别稳定在0.5和0.6左右,且用户长记忆效应变动与用户度呈正相关,Pearson相关系数分别为0.935 8和0.927 8。该工作有助于深入理解用户集群行为和信任关系的动态演化机制。
关键词 集群行为; 信任关系; 去趋势波动分析法; 赫斯特指数
基金项目 国家自然科学基金面上项目(6173248,71771152)
本文URL http://www.arocmag.com/article/01-2019-08-007.html
英文标题 Study of long-term memory in online rating behavior for trust formation
作者英文名 Guo Xinyu, Guo Qiang, Liu Jianguo
机构英文名 Research Center of Complex Systems Science,University of Shanghai for Science & Technology,Shanghai 200093,China
英文摘要 Investigating the dynamics of long-term memory in online rating behaviors is significant for understanding the evolution mechanism of collective behaviors and trust formation for online social networks. Since users are allowed to deliver ratings in many online systems, ratings can well reflect the user's opinions. This paper empirically investigated the long-term memory, measured by the detrended fluctuation analysis, in collective rating behaviors before and after the trust formation. The results for the Epinions data set show that, comparing with the null model generated by the reshuffle process, the Hurst exponent of trustors decreased 8.06% before and increased 8.43% after trust formation, which stably remained close to 0.5 in null model I and 0.6 in null model Ⅱ, suggesting that the collective rating behavior played an important role for the trust formation. Furthermore, this paper divided users into 8 groups according to the user degree and found that the correlation of the user degree and the variation of Hurst exponent, measured by the Pearson correlation coefficient, was 0.935 8 and 0.927 8 before and after trust formation respectively, reflecting a significant correlation between user degrees and collective rating behavior patterns. This work helps deeply understand the intrinsic feedback effects between collective behaviors and trust relationship.
英文关键词 collective behavior; trust formation; detrended fluctuation analysis(DFA); Hurst exponent
参考文献 查看稿件参考文献
 
收稿日期 2018/3/6
修回日期 2018/4/9
页码 2275-2278
中图分类号 TP311.1
文献标志码 A