MULTIFRACTAL DETRENDED CROSS-CORRELATION ANALYSIS OF HIGH-FREQUENCY STOCK SERIES BASED ON ENSEMBLE EMPIRICAL MODE DECOMPOSITION

Abstract

We used the multifractal detrended cross-correlation analysis (MFDCCA) method based on ensemble empirical mode decomposition (EEMD) to study the 5-min high-frequency data of two Chinese stocks and two US stocks. Using EEMD method to decompose the original high-frequency stock data can effectively reduce the interference of noise on the series, which helps to reveal the internal characteristics of the stock system and extract more accurate and rich information. We first conducted a cross-correlation test and cross-correlation coefficient analysis on the reconstructed stock data of two groups, and found that there is a cross-correlations between them. Then we used the EEMD-based MFDCCA method to analyze the cross-correlation between the data and found that there are significant cross-correlations between DJI and NASDAQ and between SSEC and SZSE. The cross-correlation of the two Chinese stocks is stronger than that of the two US stocks. The MFDCCA results of the comparison of the original series with the reconstructed series after decomposition by the EEMD method show that the reconstructed series can display more internal details of the multifractal cross-correlation metrics compared with the original series.