Proceedings of the 2015 SIAM International Conference on Data Mining 2015
DOI: 10.1137/1.9781611974010.47
|View full text |Cite
|
Sign up to set email alerts
|

Hidden Hazards: Finding Missing Nodes in Large Graph Epidemics

Abstract: Given a noisy or sampled snapshot of an infection in a large graph, can we automatically and reliably recover the truly infected yet somehow missed nodes? And, what about the seeds, the nodes from which the infection started to spread? These are important questions in diverse contexts, ranging from epidemiology to social media.In this paper, we address the problem of simultaneously recovering the missing infections and the source nodes of the epidemic given noisy data. We formulate the problem by the Minimum D… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
30
0

Year Published

2017
2017
2022
2022

Publication Types

Select...
5
4

Relationship

0
9

Authors

Journals

citations
Cited by 27 publications
(30 citation statements)
references
References 25 publications
0
30
0
Order By: Relevance
“…e closest in the area of reconstructing epidemics over time has been done in [14] where the authors use Steiner trees to infer the propagation structure. However in most real datasets, observing the cascade from the start is di cult due to time of observation [17] and missing data issues. In that context, our work only relies on the current information available to determine the source of infections.…”
Section: Related Workmentioning
confidence: 99%
“…e closest in the area of reconstructing epidemics over time has been done in [14] where the authors use Steiner trees to infer the propagation structure. However in most real datasets, observing the cascade from the start is di cult due to time of observation [17] and missing data issues. In that context, our work only relies on the current information available to determine the source of infections.…”
Section: Related Workmentioning
confidence: 99%
“…[22] studies the network completion problem, where the focus is to learn a probabilistic model that fits the observed part of a network, and then uses the model to infer missing nodes and links of the network. More specifically, [23] addresses the problem of recovering the missing infections and the source nodes of an epidemic from sampled snapshots of large graphs. The notion of graph identification is introduced in [24], which aims to infer a cleaned output network from a noisy, incomplete input graph.…”
Section: B Incomplete Graph Miningmentioning
confidence: 99%
“…The work closest to ours is the work by Sundareisan et al [3]. Similar to our problem formulation, they aim at recovering hidden infections under a non-temporal setting where the source nodes are unknown.…”
Section: Related Workmentioning
confidence: 99%
“…• We define the probabilistic cascade-reconstruction problem (Section III), which makes weaker assumptions compared to methods such as NetFill [3], thus offering more robustness. • To solve the cascade-reconstruction problem, we study the problem of sampling Steiner trees with a given set of terminals, and propose two algorithms with provable guarantees on the sampling distribution (Section IV).…”
Section: Introductionmentioning
confidence: 99%