Several network science applications involve nodal processes with dynamics dependent on the underlying graph topology that can possibly jump over discrete states. The connectivity in dynamic brain networks for instance, switches among candidate topologies, each corresponding to a different emotional state. In this context, the present work relies on limited nodal observations to perform semi-supervised tracking of dynamic processes over switching graphs. To this end, leveraging what is termed interacting multi-graph model (IMGM), a scalable online Bayesian approach is developed to track the active graph topology and dynamic nodal process. Numerical tests with synthetic and real datasets demonstrate the merits of the novel approach.
|Original language||English (US)|
|Title of host publication||2019 IEEE Data Science Workshop, DSW 2019 - Proceedings|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|Number of pages||5|
|State||Published - Jun 2019|
|Event||2019 IEEE Data Science Workshop, DSW 2019 - Minneapolis, United States|
Duration: Jun 2 2019 → Jun 5 2019
|Name||2019 IEEE Data Science Workshop, DSW 2019 - Proceedings|
|Conference||2019 IEEE Data Science Workshop, DSW 2019|
|Period||6/2/19 → 6/5/19|
Bibliographical noteFunding Information:
This paper dealt with tracking dynamic graph processes that evolve over switching graph topologies. Given observations at a subset of nodes and candidate mode-conditioned topologies, a scalable Bayesian algorithm, termed IMGM, was introduced to learn the dynamic graph processes and discrete network modes online. Numerical tests on synthetic and real data corroborated the performance of the IMGM algorithm. Acknowledgement. This work was supported by NSF grants 1442686, 1508993, and 1711471.
© 2019 IEEE.
- Bayesian tracking
- Dynamic graphs