On application of the forbidden permutations patterns methods for time series analysis

Currently, one of the key challenges in many different fields of science and engineering is the development of methods capable of distinguishing noise signals from chaotic ones. Analysis of the nature and structure of temporal data series can predict many adverse events before they occur: heart or epileptic attacks, various engine breakdowns, changes in financial markets, etc. The problem of distinguishing between signals of random nature (noise signals) and signals whose nature is determined by complex non-periodic (chaotic) dynamics, is not yet completely solved. This work was aimed to create a method of time series analysis based on forbidden permutations patterns, which will be able to distinguish the appearance of atypical dynamics. Preliminary results demonstrate the ability forbidden permutations patterns analysis method to distinguish between chaotic and noisy dynamics.