ICA Assumptions: Essential Conditions for Successful Application

To successfully apply ICA, we need to make these assumptions:

A. The values in each source signal have a symmetric distribution.

B. The source signals are statistically independent of each other.

C. The values in each source signal have a non-Gaussian distribution.

D. Each measured signal is a linear combination of the sources.

E. The mixing matrix is invertible.

F. The number of measured signals is equal to or greater than the number of source signals.

G. The sources are non-Gaussian and have different statistical properties.

H. The sources are mutually independent and have no cross-correlation.

I. The sources are stationary over the time period of observation.

J. The noise in the measured signals is additive and independent of the sources.