Multifractal Characteristics and Information Flow Analysis of Stock Markets Based on Multifractal Detrended Cross-Correlation Analysis and Transfer Entropy

Understanding cross-correlation and information flow between stocks is crucial for stock market analysis. However, traditional methods often struggle to capture financial markets’ complex and multifaceted dynamics. This paper presents a robust combination of techniques, integrating three advanced methods: Multifractal Detrended Cross-Correlation Analysis (MFDCCA), transfer entropy (TE), and complex networks. To address inherent non-stationarity and noise in financial data, we employ Ensemble Empirical Mode Decomposition (EEMD) for preprocessing, which helps reduce noise and handle non-stationary effects. The application and effectiveness of this combination of techniques are demonstrated through examples, uncovering significant multifractal properties and long-range cross correlations among the stocks studied. This combination of techniques also captures the magnitude and direction of information flow between stocks. This holistic analysis provides valuable insights for investors and policymakers, enhancing their understanding of stock market behavior and supporting better-informed portfolio decisions and risk management strategies.