Estatísticas de avalanches e intervalos interavalanches em séries temporais de sistemas complexos

Abstract

In this work, we present empirical results obtained through the analysis of time series of three different complex systems: barkhausen noise signals in magnetic systems, accelerometer signals that recorded motor activity in rodents and acoustic signals that recorded background noise in some ecosystems (soundscapes). In particular, we found that the time series of these three systems have a common feature: intermittency, all with some time interval distributed according to a power law. More precisely, in the case of Barkhausen noise, we obtained an exponential distribution for waiting times, in contrast to the already well-known power law for avalanche durations. For the accelerometer signals obtained in mice, we show that there is an alternation between a distribution of the power law form for the durations of the avalanches of motor activity and an exponential distribution for the quiet intervals. In the case of acoustic time series, we show that there is an alternation between a lognormal distribution for the intervals of acoustic activity (the sound times) and a power law for the intervals of quiet for all analyzed ecosystems.