Open AccessClassification of chaotic time series data based on the average length of close orbits
Author(s) -
Xia Heng-Chao,
Zhan Yong-Qi
Publication year - 2004
Publication title -
wuli xuebao
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.199
H-Index - 47
ISSN - 1000-3290
DOI - 10.7498/aps.53.1299
Subject(s) - chaotic , series (stratigraphy) , similarity (geometry) , waveform , recurrence plot , plot (graphics) , computer science , time series , rhythm , mean squared prediction error , lorenz system , algorithm , pattern recognition (psychology) , statistics , mathematics , artificial intelligence , physics , nonlinear system , paleontology , telecommunications , radar , quantum mechanics , acoustics , image (mathematics) , biology
Based on recurrence plot (RP) of chaotic time series, this paper develops the definition of cross recurrence plot (CRP) and the average length of close orbits, which reflects the similarity of two time series and is used to classify time series. By analysing the Lorenz signals, we conclude that the average length is decreased with the decrease of the similarity of two orbits. We have used our method and the crossprediction error in the detection of ECG rhythm. The results show that both methods work for rhythm with different waveforms, but our method works better for rhythms with similar waveforms than the method of cross_prediction error.