Reconstructing an Epidemic Over Time

We consider the problem of reconstructing an epidemic over time, or, more general, reconstructing the propagation of an activity in a network. Our input consists of a temporal network, which contains information about when two nodes interacted, and a sample of nodes that have been reported as infected. The goal is to recover the flow of the spread, including discovering the starting nodes, and identifying other likely-infected nodes that are not reported. The problem we consider has multiple applications, from public health to social media and viral marketing purposes. Previous work explicitly factor-in many unrealistic assumptions: it is assumed that (a) the underlying network does not change;(b) we have access to perfect noise-free data; or (c) we know the exact propagation model. In contrast, we avoid these simplifications: we take into account the temporal network, we require only a small sample of reported infections, and we do not make any restrictive assumptions about the propagation model. We develop CulT, a scalable and effective algorithm to reconstruct epidemics that is also suited for online settings. CulT works by formulating the problem as that of a temporal Steiner-tree computation, for which we design a fast algorithm leveraging the specific problem structure. We demonstrate the efficacy of the proposed approach through extensive experiments on diverse datasets.

[1]  Krishna P. Gummadi,et al.  On the evolution of user interaction in Facebook , 2009, WOSN '09.

[2]  Michalis Faloutsos,et al.  Threshold conditions for arbitrary cascade models on arbitrary networks , 2011, 2011 IEEE 11th International Conference on Data Mining.

[3]  Jure Leskovec,et al.  Correcting for missing data in information cascades , 2011, WSDM '11.

[4]  Sudipto Guha,et al.  Approximation algorithms for directed Steiner problems , 1999, SODA '98.

[5]  Jilles Vreeken,et al.  Hidden Hazards: Finding Missing Nodes in Large Graph Epidemics , 2015, SDM.

[6]  P. Holme,et al.  Predicting and controlling infectious disease epidemics using temporal networks , 2013, F1000prime reports.

[7]  Othon Michail,et al.  An Introduction to Temporal Graphs: An Algorithmic Perspective* , 2015, Internet Math..

[8]  Yao Zhang,et al.  Scalable Vaccine Distribution in Large Graphs given Uncertain Data , 2014, CIKM.

[9]  Christos Faloutsos,et al.  Spotting Culprits in Epidemics: How Many and Which Ones? , 2012, 2012 IEEE 12th International Conference on Data Mining.

[10]  Carl Kingsford,et al.  Diffusion Archaeology for Diffusion Progression History Reconstruction , 2014, ICDM.

[11]  Hongyuan Zha,et al.  Back to the Past: Source Identification in Diffusion Networks from Partially Observed Cascades , 2015, AISTATS.

[12]  Jari Saramäki,et al.  Temporal Networks , 2011, Encyclopedia of Social Network Analysis and Mining.

[13]  Ada Wai-Chee Fu,et al.  Minimum Spanning Trees in Temporal Graphs , 2015, SIGMOD Conference.

[14]  Dimitrios Gunopulos,et al.  Finding effectors in social networks , 2010, KDD.

[15]  Devavrat Shah,et al.  Rumors in a Network: Who's the Culprit? , 2009, IEEE Transactions on Information Theory.

[16]  Hanghang Tong,et al.  Full diffusion history reconstruction in networks , 2015, 2015 IEEE International Conference on Big Data (Big Data).