Skip to main content

Advances, Systems and Applications

Table 1 Advantages and disadvantages of several common sequence decomposition methods

From: Time series forecasting model for non-stationary series pattern extraction using deep learning and GARCH modeling

Series Decomposition Methods

Method

Advantage

Disadvantage

VMD

Featuring rigorous theoretical derivation, widespread application, and relatively low computational complexity [34].

the decomposed results are heavily restricted to the selection of the penalty parameter a and the number of sub-modes K [35].

EMD

Able to decompose signals adaptively [36].

mode mixing, end effect, poor noise immunity [37].

EEMD

Has strong adaptability, effectively overcoming the phenomenon of mode mixing [38].

its reconstruction error is large and its integrity is poor [39].

CEEMD

Ensure decomposition effectiveness, while reducing reconstruction errors caused by white noise [40].

There is a loss of information for high frequency components [41].

CEEMDAN

Successfully resolved the issue of white noise transmission from high frequency to low frequency [42].

There exist noise after decomposing complex sequential data [43].