Politi Antonio
Institute for Complex Systems and Mathematical Biology, SUPA, University of Aberdeen, AB24 3UE, Aberdeen, United Kingdom.
Phys Rev Lett. 2017 Apr 7;118(14):144101. doi: 10.1103/PhysRevLett.118.144101.
A powerful approach is proposed for the characterization of chaotic signals. It is based on the combined use of two classes of indicators: (i) the probability of suitable symbolic sequences (obtained from the ordinal patterns of the corresponding time series); (ii) the width of the corresponding cylinder sets. This way, much information can be extracted and used to quantify the complexity of a given signal. As an example of the potentiality of the method, I introduce a modified permutation entropy which allows for quantitative estimates of the Kolmogorov-Sinai entropy in hyperchaotic models, where other methods would be unpractical. As a by-product, estimates of the fractal dimension of the underlying attractors are possible as well.
提出了一种用于表征混沌信号的有效方法。它基于两类指标的联合使用：（i）合适符号序列的概率（从相应时间序列的序数模式获得）；（ii）相应柱集的宽度。通过这种方式，可以提取大量信息并用于量化给定信号的复杂性。作为该方法潜力的一个例子，我引入了一种改进的排列熵，它允许在超混沌模型中对柯尔莫哥洛夫-西奈熵进行定量估计，而其他方法在此处并不实用。作为一个副产品，也可以对基础吸引子的分形维数进行估计。
