Fast principal component analysis and data whitening algorithms

In this paper, we propose an adaptive implementation of a fast-convergent algorithm for principal component extraction. Our approach consists of first estimating a basis of the principal subspace through the use of OPAST algorithm. The obtained basis is then fed to a second process where at each iteration one or several Givens transformations are applied to estimate the principal components. Later on, the proposed PCA algorithm is used to derive a fast data whitening solution that overcomes the existing ones of similar complexity order. Simulation results support the high performance of our algorithms in terms of accuracy and speed of convergence.