Condensation phenomena in fat-tailed distributions: a characterization by means of an order parameter
arXiv:1309.7795
Abstract
Condensation phenomena are ubiquitous in nature and are found in condensed matter, disordered systems, networks, finance, etc. In the present work we investigate one of the best frameworks in which condensation phenomena take place, namely, the sum of independent and fat-tailed distributed random variables. For large deviations of the sum, this system undergoes a phase transition and shifts from a democratic phase to a condensed phase, where a single variable (the condensate) carries a finite fraction of the sum. This phenomenon yields the failure of the standard results of the Large Deviation Theory. In this work we exploit the Density Functional Method to overcome the limitation of the Large Deviation Theory and characterize the condensation transition in terms of an order parameter, i.e. the Inverse Participation Ratio (IPR). This procedure leads us to investigate the system in the large-deviation regime where both the sum and the IPR are constrained, observing new phase transitions. As a sample application, the case of condensation phenomena in financial time-series is briefly discussed.
arXiv comment23 pages, 5 figures
References in corpus (8)
- Bose-Einstein condensation in complex networks
- Universality Classes for Extreme Value Statistics
- Large Deviations of Extreme Eigenvalues of Random Matrices
- "Universal" Distribution of Inter-Earthquake Times Explained
- The Nature of the Condensate in Mass Transport Models
- Phase Transitions in the Distribution of Bipartite Entanglement of a Random Pure State
- Rain: Relaxations in the sky
- Constraint driven condensation in large fluctuations of linear statistics