Open AccessComparative analysis of three distribution entropy methods for chaos recognition
Author(s) -
Xiangjian Zeng,
Li Wan,
Hui Liu
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1732/1/012060
Subject(s) - chaotic , entropy (arrow of time) , logistic map , approximate entropy , mathematics , algorithm , sequence (biology) , pattern recognition (psychology) , computer science , artificial intelligence , statistical physics , physics , quantum mechanics , biology , genetics
Distribution entropy (DistEn) is an effective index to measure the complexity of time series. In this paper, the moving distribution entropy (M-DistEn) method is proposed by combining the distribution entropy method with the moving window technology. The moving cut data-distribution entropy (MC-DistEn) method is proposed by combining the DistEn with the moving removal window. Based on the M-DistEn method, moving weighted distribution entropy (MW-DistEn) is proposed. These three distribution entropy methods are used to identify the chaotic state of mixed Logistic map sequences with different parameters. The results show that the M-DistEn method only recognizes the first three mixed states of the sequence, and it cannot accurately recognize the last two chaotic states of the sequence, so it has certain limitations. The MC-DistEn method can accurately identify the four different chaotic states of the sequence, but the DistEn value greatly fluctuates due to the influence of the size of the moving removal window, and the position judgment of sequence state change is not accurate enough. The MW-DistEn method not only accurately recognizes different chaotic states, but also is more accurate in judging the position of sequence state changes and more stable in DistEn value than the M-DistEn method and MC-DistEn method, thereby it has a good application prospect.