EFFECT OF WINDOW LENGTH WHEN SMOOTHING WITH SINGULAR SPECTRUM ANALYSIS TECHNIQUE IN RUNNING DATA

Kazuhiro Ishimura; Shinji Sakurai

EFFECT OF WINDOW LENGTH WHEN SMOOTHING WITH SINGULAR SPECTRUM ANALYSIS TECHNIQUE IN RUNNING DATA

##article.authors##

Kazuhiro Ishimura
Shinji Sakurai

##semicolon##

smoothing##common.commaListSeparator## SSA##common.commaListSeparator## running kinematics filter##common.commaListSeparator## optimum cutoff frequency

##article.abstract##

The purpose of this study was to demonstrate the effect of window length on running kinematic data in comparison to data smoothed using a digital Butterworth low pass filter. The raw marker trajectories were smoothed using a digital Butterworth filter with an optimum cutoff frequency and using the SSA technique with window lengths of L = n/2, n/5, and n/10 (n = 220). Data smoothing using SSA parameters of L = n/10, r = 2 produced similar results to the Butterworth low pass filter with an optimal cutoff frequency of 13 Hz. In conclusion, a window length of L = n/10 is recommended for running kinematics research, while window lengths of L = n/2 should be avoided.