Financial time series are nonlinear and non-stationary, and they are inherently noisy, it is adjudged the most challenging problem for financial time series. For the nonlinear and non-stationary time series, the empirical mode decomposition can decompose them into several intrinsic mode functions (IMFs) adaptively which is part of Hilbert-Huang transform (HHT). Thus, this paper proposed the novel improved HHT based financial time series de-noising method using compositional data techniques. Utilizing the simulation study and pragmatic research of forecasting based on Gold Closing Price. Four popular denoising methods, i.e. Wavelet, EMD-based hard and soft thresholding, EMD-based Savitzky-Golay filter, are also performed for comparison purpose. Both the simulation and empirical results suggest that the proposed method is a valid and practical value for denosing and prediction of the financial time series.
Paper
Full text
A New HHT-Based Denoising Algorithm for Financial Time Series Data Mining
Semantic Scholar · Computer Science · 2019
Abstract
Financial time series are nonlinear and non-stationary, and they are inherently noisy, it is adjudged the most challenging problem for financial time series. For the nonlinear and non-stationary time series, the empirical mode decomposition can decompose them into several intrinsic mode functions (IMFs) adaptively which is part of Hilbert-Huang transform (HHT). Thus, this paper proposed the novel improved HHT based financial time series de-noising method using compositional data techniques. Utilizing the simulation study and pragmatic research of forecasting based on Gold Closing Price. Four popular denoising methods, i.e. Wavelet, EMD-based hard and soft thresholding, EMD-based Savitzky-Golay filter, are also performed for comparison purpose. Both the simulation and empirical results suggest that the proposed method is a valid and practical value for denosing and prediction of the financial time series.