In the field of network theory, the so called “explosive percolation” phenomenon has recently raised great interest, in which a giant cluster transition is obtained by the addition of a negligibly small number of links (as opposed to erdos-renyi classical random networks). We describe a similar behaviour in networks obtained from a correlation matrix of randomly generated vectors, both in the limit of high correlation and anti-correlation. This phenomenon is relevant in many practical applications, from financial to high-throughput genomic data analysis, in which time correlation is widely applied to characterize the global and local dynamics of the system.
Titolo: | Explosive percolation in correlation-based networks |
Autore/i: | REMONDINI, DANIEL; MENICHETTI, GIULIA; SERVIZI, GRAZIANO; CASTELLANI, GASTONE |
Autore/i Unibo: | |
Anno: | 2010 |
Titolo del libro: | Atti del XCVI congresso nazionale della Società Italiana di Fisica |
Pagina iniziale: | 90 |
Pagina finale: | 90 |
Abstract: | In the field of network theory, the so called “explosive percolation” phenomenon has recently raised great interest, in which a giant cluster transition is obtained by the addition of a negligibly small number of links (as opposed to erdos-renyi classical random networks). We describe a similar behaviour in networks obtained from a correlation matrix of randomly generated vectors, both in the limit of high correlation and anti-correlation. This phenomenon is relevant in many practical applications, from financial to high-throughput genomic data analysis, in which time correlation is widely applied to characterize the global and local dynamics of the system. |
Data prodotto definitivo in UGOV: | 18-giu-2013 |
Appare nelle tipologie: | 4.02 Riassunto (Abstract) |