Composite Multiscale Partial Cross-Sample Entropy Analysis for Quantifying Intrinsic Similarity of Two Time Series...

Area of Science:

• Time series analysis
• Complexity science
• Financial econometrics

Background:

• Quantifying time series similarity is challenging, especially when external factors influence the data.
• Existing methods may not fully isolate intrinsic relationships between series.

Purpose of the Study:

• To propose a novel method, composite multiscale partial cross-sample entropy (CMPCSE), for measuring intrinsic similarity.
• To validate CMPCSE's effectiveness in removing external influences.

Main Methods:

• Development of the composite multiscale partial cross-sample entropy (CMPCSE) algorithm.
• Application of CMPCSE to artificial datasets for validation.
• Analysis of Shanghai and Shenzhen stock indices, controlling for the Hang Seng Index.

Main Results:

• CMPCSE accurately measures intrinsic cross-sample entropy by removing third-party time series effects.
• Application to financial indices revealed stronger intrinsic similarity between SSEC and SZSE than previously shown.
• CMPCSE demonstrated superior performance compared to composite multiscale cross-sample entropy.

Conclusions:

• CMPCSE is a robust tool for assessing the intrinsic similarity of time series.
• The method effectively isolates underlying relationships obscured by common external factors.
• This technique offers valuable insights into financial market dynamics.