Distortion Impact on Low-Dimensional Manifold Recovery of High-Dimensional Data

  • Alaa Abd El Hakeem ,
  • Motaz El-Saban

Taibah University International Conference on Computing and Information Technology |

In this paper, we investigate the effect of various kinds of distortions on the performance of subspace recovery using Principal Component Analysis (PCA). We verify that while PCA demonstrates relative good stability characteristics in presence of mild distortions, it suffers from major shortcomings to gross corruption of input observations, even if these corruptions are sparse. We present a performance evaluation study for the performance of the classical PCA under observation distortions and compare it with the performance of the Robust Principal Component Analysis (RPCA) approach. We verify the effectiveness of RPCA in lower-dimensional subspace recovery under different kinds of distortions.