ABSTRACT
In the frequency domain the basic idea is to cut off the high frequency components of a given line.
Fourier transform
and
wavelet transform
are two commonly used tools. More recently, a data-driven decomposition
method, called
empirical mode decomposition
(EMD
), has been developed (Huang et al., 1998) and used for smoothing data to obtain trends (Li et al., 2004). In the space domain two approaches are widely used:
moving averaging
and
curve fitting
. Three types of curve fitting can be distinguished:
least-squares fitting
,
exact fitting,
and
energy-balancing fitting
.