Sampling Properties of color Independent Component Analysis

Area of Science:

• Statistics
• Signal Processing
• Data Science

Background:

• Independent Component Analysis (ICA) is a data-driven method for blind source extraction.
• Existing ICA methods lack rigorous statistical inference investigation in the statistics literature.

Purpose of the Study:

• To investigate the statistical sampling properties of the colorICA (cICA) method.
• To bridge the gap in statistical inference for ICA methods.

Main Methods:

• Utilized parametric time series models in the frequency domain to incorporate source correlation.
• Analyzed the consistency and asymptotic normality of cICA estimates.

Main Results:

• Established theoretical consistency and asymptotic normality for cICA estimates.
• Demonstrated superior numerical performance of cICA over existing alternatives.
• Validated asymptotic properties through simulation studies.

Conclusions:

• The cICA method provides a statistically sound approach for blind source extraction.
• The established properties enable reliable statistical inference in signal and image processing.
• cICA offers a robust alternative for complex source separation tasks.