A method for segmentation of switching dynamic modes in time series.

Area of Science:

• Complex systems analysis
• Time series analysis
• Nonlinear dynamics

Background:

• The Annealed Competition of Experts (ACE) algorithm identifies switching dynamics in time series.
• Performance of ACE is sensitive to embedding dimension and time delay selection.
• Systematic methods are needed to optimize ACE parameters.

Purpose of the Study:

• To improve the accuracy and robustness of the ACE algorithm.
• To develop systematic approaches for parameter selection in ACE.
• To validate the enhanced ACE method on physiological time series data.

Main Methods:

• Utilized mutual information and false nearest neighbor for optimal embedding dimension and time delay selection.
• Incorporated deterministic annealing and phase space closeness measure to refine the ACE algorithm.
• Applied the enhanced ACE method to rat heart rate data under varying autonomic blockade conditions.

Main Results:

• The improved ACE method accurately determines the location of switching dynamic modes in time series.
• Parameter selection using mutual information and false nearest neighbor enhanced ACE performance.
• The enhanced ACE algorithm demonstrated superior accuracy in segmenting dynamic mode changes compared to the original ACE.

Conclusions:

• The enhanced ACE algorithm provides a more accurate and reliable tool for analyzing switching dynamics in time series.
• Systematic parameter optimization is crucial for the effective application of the ACE method.
• The improved ACE method shows significant potential for applications in physiological signal analysis and other complex systems.