Effects of controlling parameter on symbolic nonlinear complexity detection

Abstract: Symbolic transformation, a coarse-graining process, is a crucial prerequisite for and has evidential influence to the symbolic time series analysis. We employ Shannon entropy for a parameter-dependent symbolization, KW (Kurths-Wessel) symbolic method, to test the effects of controlling parameter on its symbolic nonlinear complexity detection. Two chaotic models, logistic and Henon series, and heartbeats of CHF (Congestive Heart Failure) patients, healthy young and elderly subjects from PhysioNet are applied to test the KW symbolic entropy. The complexity-loss theory about aging and diseases in heart rates is validated and reasons that may account for some paradoxes in nonlinear analysis are discussed. Tests results suggest that due to different structural or dynamical properties of different nonlinear systems, controlling parameter of the KW symbolization should be adjusted accordingly to have reliable symbolic nonlinear analysis.