Instability of Financial Time Series Revealed by Irreversibility Analysis

Area of Science:

• Quantitative Finance
• Complex Systems Analysis
• Time Series Analysis

Background:

• The 2008 global economic crisis heightened the need for robust financial instability detection.
• Time-series analysis is a key tool for understanding financial market dynamics.
• Existing methods often struggle to capture complex temporal patterns and anomalies.

Purpose of the Study:

• To introduce a novel time-series analysis method for detecting financial instabilities.
• To evaluate the efficacy of the Kullback-Leibler Divergence (KLD) metric integrated with a sliding window technique.
• To demonstrate the method's superiority over traditional approaches.

Main Methods:

• Global financial time series (2004-2022) were analyzed.
• Data were transformed into return rate series and complex networks via the directed horizontal visibility graph (DHVG) algorithm.
• The Kullback-Leibler Divergence (KLD) metric was applied using a sliding window approach.

Main Results:

• The KLD method successfully identified specific financial market incidents and correlated them with economic events.
• Retrospective analysis and real-time monitoring confirmed the method's effectiveness.
• KLD outperformed traditional metrics and econometric methods in capturing sequential information and detecting anomalies.

Conclusions:

• The proposed KLD-based sliding window method is a powerful tool for detecting financial instabilities.
• This technique offers enhanced anomaly detection capabilities compared to conventional statistical and econometric models.
• The method's versatility allows for application beyond financial markets to other time-series data.