A Dynamic Extraction Model for Susceptible Users Based on DSSI Theory

doi:10.11897/SP.J.1016.2017.02812

SIC OpenIR

	A Dynamic Extraction Model for Susceptible Users Based on DSSI Theory
	Li, Ling 1,2; Liu, Min 1; Cheng, Guo-Qing 2
	2017
发表期刊	Jisuanji Xuebao/Chinese Journal of Computers
ISSN	02544164
卷号	40 期号:12 页码:2812-2826
摘要	Fast growing social networks have been integrated into people's daily lives and play an important role, which makes more and more academics study the social networks from different perspectives. The research of information diffusion over online social networks can help users to obtain information, enterprises to promote product, politicians to regulate public opinion, and is with the significant value in theory and application. In the recent years, although there have been a number of significant advancements on information diffusion, most of them have been mainly focusing on optimization algorithm for user extraction, or evolution equations for behavior law. No attempts have been made to quantify user susceptibility through the forwarding action, and dynamically extract susceptible users. In fact, at different times, the influence of users on information diffusion cannot be exactly the same due to the uncertainty and complexity of social networks. In this sense, dynamic analysis and study of user forwarding action is of great importance, and is precisely what we do in this paper. To address dynamically the business information diffusion problem over online social networks, the randomness and uncertainties of user forwarding action are first analyzed, and then a novel dynamic extraction model for the susceptible users is presented, which is based on Universal Generating Function(UGF) method and Discrete Stress-Strength Interference (DSSI) theory. In the model, the random forwarding action of the user is firstly quantified as Node Susceptibility(NS), and NS is relevant to two random variables of information receiving Xnmtand forwarding Ynmt. Then, according to UGF method and DSSI theory, the values of NS in regard to different kinds of information at different periods are obtained by deriving the probability distributions of Xnmtand Ynmt, and UGFs of Xnmtand Ynmt. Finally, the susceptible users are extracted based on dynamic order of the values of NS. The decision results of the model can effectively address the following three issues: (1) the most susceptible users; (2) the kinds of information that they are most susceptible to; and (3) the period when they are most susceptible. The answers to these three questions can provide theoretical basis for making effective strategy of information diffusion, and the decision results can be updated dynamically with the observation parameters. A case study of online group buying website illustrates the feasibility and practicality of the proposed model. Further, based on the same experimental data set, the proposed model is compared with Influence-Susceptibility-Cynical(ISC) model in literature, and different susceptible users are extracted based on ISC model and our model. The results show that the susceptible users extracted in these two models are roughly the same, and the consistency is more than 70 percent. The consistency also illustrates the validity of our model to some extent. On the other hand, the difference between user extractions in different models is analyzed from both theoretical and practical perspectives. Since the quantification of user susceptibility in our model is based on the statistics characteristics of observation parameters, it is concluded that our model is more scientific and reasonable in quantifying user susceptibility. © 2017, Science Press. All right reserved.
DOI	10.11897/SP.J.1016.2017.02812
EI入藏号	20182305288945
EI主题词	Social networking (online)
EI分类号	723 Computer Software, Data Handling and Applications - 723.5 Computer Applications - 802.3 Chemical Operations - 901.4 Impact of Technology on Society - 922.1 Probability Theory
引用统计
文献类型	期刊论文
条目标识符	http://ir.sic.ac.cn/handle/331005/25922
专题	中国科学院上海硅酸盐研究所
作者单位	1.College of Electronics and Information Engineering, Tongji University, Shanghai; 201804, China; 2.Department of Information Engineering, Jingdezhen Ceramic Institute, Jingdezhen; Jiangxi; 333403, China
推荐引用方式 GB/T 7714	Li, Ling,Liu, Min,Cheng, Guo-Qing. A Dynamic Extraction Model for Susceptible Users Based on DSSI Theory[J]. Jisuanji Xuebao/Chinese Journal of Computers,2017,40(12):2812-2826.
APA	Li, Ling,Liu, Min,&Cheng, Guo-Qing.(2017).A Dynamic Extraction Model for Susceptible Users Based on DSSI Theory.Jisuanji Xuebao/Chinese Journal of Computers,40(12),2812-2826.
MLA	Li, Ling,et al."A Dynamic Extraction Model for Susceptible Users Based on DSSI Theory".Jisuanji Xuebao/Chinese Journal of Computers 40.12(2017):2812-2826.