next up previous
Next: Preprocessing of the data Up: Algorithms for ICA Previous: Algorithms for ICA


After choosing one of the principles of estimation for ICA discussed in Section 4, one needs a practical method for its implementation. Usually, this means that after choosing an objective function for ICA, we need to decide how to optimize it. In this Section, we shall discuss the optimization problem.

The statistical properties of the ICA method depend only on the objective functions used. Thus, the statistical properties of the objective functions were treated in Section 4. In this Section, we shall compare the different algorithms mainly on the basis of the stability, convergence speed, and memory requirements.

Aapo Hyvarinen