Name
Abstract | ||
|---|---|---|
The forwarding behavior of each individual at microscopic level determines the final morphology of macroscopic cascades. Fully grasping the inner relationship between forwarding events is essential for understanding the intrinsic diffusion mechanisms, which has proven to offer decisive references in many applications, e.g., viral marketing and innovation diffusion. Conventional approaches generally constrain the occurrence of forwarding events to a specific priori based on expert experience. However, excessive parameterization will limit the learning ability of these models. As a typical stochastic process that can characterize the mutual excitation phenomenon, the Hawkes process merely assumes that the influence between users must satisfy the mode of "excitation first, then decay". In this paper, we propose the Coupled Self-Exciting Process (CoSEP) to achieve a generalized stochastic process for information diffusion prediction. Concretely, we design a coupling strategy based on the shifted Bi-LSTM, which simultaneously considers the instantaneous excitation generated by history and the subsequent evolution while modeling the change of intensity function corresponding to each participant. Extensive experiments on real-world datasets, Sina Weibo and Open Academic Graph 2.0, demonstrate that the proposed CoSEP significantly outperforms state-of-the-art approaches, suggesting the effectiveness of the coupled self-exciting process for social applications. |
Year | DOI | Venue |
|---|---|---|
2021 | 10.1109/ISPA-BDCloud-SocialCom-SustainCom52081.2021.00190 | 19TH IEEE INTERNATIONAL SYMPOSIUM ON PARALLEL AND DISTRIBUTED PROCESSING WITH APPLICATIONS (ISPA/BDCLOUD/SOCIALCOM/SUSTAINCOM 2021) |
Keywords | DocType | ISSN |
information diffusion, user activation, deep learning, Hawkes process | Conference | 2158-9178 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Yuyang Liu | 1 | 4 | 3.43 |
| Yuxiang Ma | 2 | 0 | 1.35 |
| Junruo Gao | 3 | 0 | 0.68 |
| Zefang Zhao | 4 | 0 | 0.34 |
| Jun Li | 5 | 0 | 1.01 |