Messager, Antoine, Parisis, Georgios, Kiss, István Z, Harper, Robert, Tee, Phil and Berthouze, Luc (2019) Functional topology inference from network events. IFIP/IEEE International Symposium on Integrated Network Management. Intelligent Management for the Next Wave of Cyber and Social Networks, Washington DC, USA, 8-12 April 2019. Published in: 2019 IFIP/IEEE Symposium on Integrated Network and Service Management (IM). Institute of Electrical and Electronics Engineers ISBN 9783903176157
![]() |
PDF
- Accepted Version
Download (245kB) |
Abstract
In this paper we present a novel approach for inferring functional connectivity within a large-scale network from time series of emitted node events. We do so under the following constraints: (a) non-stationarity of the underlying connectivity, (b) sparsity of the time-series of events, and (c) absence of an explicit model describing how events propagate through the network. We develop an inference method whose output is an undirected weighted network, where the weight of an edge between two nodes denotes the probability of these nodes being functionally connected. Two nodes are assumed to be functionally connected if they show significantly more coincident or short-lagged events than randomly picked pairs of nodes with similar levels of activity. We develop a model of time-varying connectivity whose parameters are determined by maximising the model’s predictive power from one time window to the next. We assess the accuracy, efficiency and scalability of our method on a real dataset of network events spanning multiple months.
Item Type: | Conference Proceedings |
---|---|
Keywords: | network management, network events, topology inference, functional connectivity, machine learning |
Schools and Departments: | School of Engineering and Informatics > Informatics School of Mathematical and Physical Sciences > Mathematics |
Research Centres and Groups: | Sussex Neuroscience |
Subjects: | Q Science > QA Mathematics > QA0075 Electronic computers. Computer science T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5101 Telecommunication > TK5105.5 Computer networks |
Depositing User: | Luc Berthouze |
Date Deposited: | 16 Jan 2019 14:19 |
Last Modified: | 28 May 2019 09:53 |
URI: | http://sro.sussex.ac.uk/id/eprint/81305 |
View download statistics for this item
📧 Request an updateProject Name | Sussex Project Number | Funder | Funder Ref |
---|---|---|---|
A fast method for calculating the proximity matrix in a large-scale dynamic network | G1742 | MOOGSOFT INC | Agreement dated 17 December 2014 |