ABSTRACT
CONTENTS 5.1 Principal Component Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 5.2 Consensus of Networked Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316 5.3 Adaptive Regulation for Hammerstein and Wiener Systems . . . . . . . . . 324 5.4 Convergence of Distributed Randomized PageRank Algorithms . . . . . 337 5.5 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
5.1 Principal Component Analysis In a practical system there may be a large amount of variables involved, but the variables may not be equally important. The principal component analysis (PCA) proposed by Pearson aims at estimating eigenvectors of a symmetric matrix in the decreasing order of importance, i.e., first to select the most important factors and then the less important factors by using linear transformations acting on the variables.
