1. Identifying the Real Influentials at Nonexplicit-Relationship Online Platforms
- Author
-
Ke Zeng, Lifang Li, Lingxi Li, and Xiao Wang
- Subjects
Sustainable development ,Measure (data warehouse) ,business.industry ,Computer science ,media_common.quotation_subject ,Stability (learning theory) ,02 engineering and technology ,Recommender system ,Data science ,Digital media ,Human-Computer Interaction ,Promotion (rank) ,Viral marketing ,020204 information systems ,Modeling and Simulation ,0202 electrical engineering, electronic engineering, information engineering ,020201 artificial intelligence & image processing ,The Internet ,business ,Social Sciences (miscellaneous) ,media_common - Abstract
The measurement of influence on online platforms has been an important issue for various applications, including viral marketing, recommender systems, and the Internet celebrity economy. Generally, the citation frequency, mention frequency, and in-degree of users are the three major criteria for evaluating online influence in existed studies. However, some online media platforms neither provide social networking functions nor support social relationship labeling, making it infeasible to measure the user influence via the above three criteria. Such platforms can be named nonexplicit-relationship platforms. In this article, we propose three new criteria, explicit conversion rate (ER), frequency of promotion (FP), and average participation density (APD), and design a novel algorithm to effectively calculate and evaluate users’ influence on these platforms. The stability and sustainability of user influence are evaluated to distinguish the real influentials from the disguised ones, while the latter usually appears for temporary commercial advertisement purposes. The experiments proved the effectiveness of the proposed criteria and the algorithm in determining influentials’ influence, as well as the corresponding stability and sustainability.
- Published
- 2020
- Full Text
- View/download PDF