An Assessment of Fractal Characterization Methods for 1/f Processes with Application to the Analysis of Stride Interval Time Series

摘要: The time evolution and complex interactions of many nonlinear systems, such as in the human body, result fractal types parameter outcomes that exhibit self similarity over long scales by a power law the frequency spectrum S(f) = 1/f. scaling exponent can be interpreted degree characteristic thus "biomarker" relative health decline. This thesis presents thorough numerical analysis characterization techniques with specific consideration given to experimentally measured gait stride interval series. ideal signals generated are constrained under varying lengths biases indicative range physiologically conceivable signals. is complement previous investigations characteristics healthy pathological series, which this study compared. comparative experimental applications provide basis for determining an appropriate robust method measuring comparing meaningful biomarker, spectral index. In constraints applications, significant drawbacks proposed domain methods noted, it concluded time-scale wavelet reasonably consistent accurate biomarker technique these