Research Presentation: Estimating Diffusion Degree on Graph Stream Generated from Social and Web Networks

Jun 17, 2024 · 1 min read
Image credit: Unsplash
Data stream generated from different Web 2.0 applications may contains data which is best described by graphs. Graph streams thus generated show big data characters, including volume and velocity. The challenges of graph stream algorithms are twofold; each edge needs to be processed only once (due to velocity), and it needs to work on highly constrained memory (due to volume). Diffusion degree, a measure of node centrality, can be calculated for static graphs using a single Breadth-First Search (BFS). However, tracking Diffusion Degree in a graph stream is nontrivial. This paper proposes an estimator for diffusion degree for graph streams, which can be used to extract top-$k$ influencing nodes for viral marketing in social networks. Comparative experiments show that the proposed graph stream algorithm is equivalent to or better than the exact diffusion degree-based algorithm.
Jun 17, 2024 — Jun 21, 2024

Tampere University


Click on the Slides button above to view the slides in new tab. Below are the embedded Slides.
ICWE 2024 Presentation by Suman Kundu
Suman Kundu
Assistant Professor of Computer Science and Engineering
My research interests include social network analysis, network data science, streaming algorithms, big data, granular computing, soft computing, fuzzy and rough sets.