Recurrence entropy and financial crashes

Abstract

Entropy is one of the most frequently and effectively used measure of the complexity of systems of various nature. And if the Shannon's canonical entropy is more a measure of the randomness of the system, then the approximate, sample, permutation and other new type entropy that have appeared recently, exploiting the Shannon entropy form have allowed us to quantify the complexity of the systems in question using fast and efficient algorithms. For the first time, a new type of recurrence entropy is used to analyze the dynamics of financial time series under crashes conditions. It is shown that recurrent entropy can be used as the indicator-predictor of financial crashes.