next up previous
Next: Kurtosis Up: Principles of ICA estimation Previous: ``Nongaussian is independent''

Measures of nongaussianity

To use nongaussianity in ICA estimation, we must have a quantitative measure of nongaussianity of a random variable, say y. To simplify things, let us assume that y is centered (zero-mean) and has variance equal to one. Actually, one of the functions of preprocessing in ICA algorithms, to be covered in Section 5, is to make this simplification possible.


Aapo Hyvarinen