Abstract:The convergence of Oja's principal component analysis (PCA) learning algorithms is a difficult topic for direct study and analysis. Traditionally, the convergence of these algorithms is indirectly analyzed via certain deterministic continuous time (DCT) systems. Such a method will require the learning rate to converge to zero, which is not a reasonable requirement to impose in many practical applications. Recently, deterministic discrete time (DDT) systems have been proposed instead to indirectly interpret the dynamics of the learning algorithms. Unlike DCT systems, DDT systems allow learning rates to be constant (which can be a nonzero). This paper will provide some important results relating to the convergence of a DDT system of Oja's PCA learning algorithm. It has the following contributions: 1) A number of invariant sets are obtained, based on which we can show that any trajectory starting from a point in the invariant set will remain in the set forever. Thus, the nondivergence of the trajectories is guaranteed. 2) The convergence of the DDT system is analyzed rigorously. It is proven, in the paper, that almost all trajectories of the system starting from points in an invariant set will converge exponentially to the unit eigenvector associated with the largest eigenvalue of the correlation matrix. In addition, exponential convergence rate are obtained, providing useful guidelines for the selection of fast convergence learning rate. 3) Since the trajectories may diverge, the careful choice of initial vectors is an important issue. This paper suggests to use the domain of unit hyper sphere as initial vectors to guarantee convergence. 4) Simulation results will be furnished to illustrate the theoretical results achieved.

Convergence of Oja's online principal component flow

Diffusion approximations of Oja's online principal component analysis

Convergence analysis of a deterministic discrete time system of Oja's PCA learning algorithm.

On the optimality of the Oja's algorithm for online PCA

Global convergence of Oja's PCA learning algorithm with a non-zero-approaching adaptive learning rate

ODE-Inspired Analysis for the Biological Version of Oja's Rule in Solving Streaming PCA

Nearly optimal stochastic approximation for online principal subspace estimation

Near-optimal stochastic approximation for online principal component estimation

Global Convergence of Oja's Subspace Algorithm for Principal Component Extraction

Online Generalized Eigenvectors Extraction Via A Fixed-Point Approach

A Riemannian Geometry Underlying Stochastic Algorithm for Adaptive Principal Component Analysis

A note on semi-groups of stochastic gradient descent and online principal component analysis

Dynamical behavior of Oja PCA model for non-symmetric covariance matrix

Semi-groups of Stochastic Gradient Descent and Online Principal Component Analysis: Properties and Diffusion Approximations

Fast and Strong Convergence of Online Learning Algorithms.

Online Learning for Sparse PCA in High Dimensions: Exact Dynamics and Phase Transitions

Convergence analysis of deterministic discrete time system of a unified self-stabilizing algorithm for PCA and MCA.

Global Convergence of a PCA Learning Algorithm with a Constant Learning Rate

Global Convergence Analysis of a Discrete Time Nonnegative ICA Algorithm

Global convergence analysis of a PCA learning algorithm

A Robust And Globally Convergent Pca Learning Algorithm