Recursive algorithms for principal component extraction

Abstract

Two new on-line recursive algorithms, namely, the Jacobi recursive principal component algorithm (JRPCA) and the Gauss–Seidel recursive principal component algorithm (GRPCA), are introduced for the computation of principal components of a slowly varying non-stationary vector stochastic process. By using these algorithms, the principal components can be adaptively estimated. The speed of convergence of the proposed algorithms is also discussed. Simulation results show that the proposed algorithms have a faster speed of convergence and a better adaptivity when compared to other existing methods.