Low‐rank isomap algorithm

Isomap is a well-known nonlinear dimensionality reduction method that highly suffers from computational complexity. Its complexity mainly arises two stages; a) embedding full graph on the data in ambient space, and b) complete eigenvalue decomposition. Although of graphing stage has been investigated by processing methods, decomposition remains bottleneck problem. In this paper, we propose Low-Rank (LRI) algorithm introducing projection operator embedded space to low-rank latent facilitate applying partial This approach leads reducing linear order while preserving structural information during process as long number observations extensively larger than space. The superiority LRI compared some state-of-art algorithms experimentally verified facial image clustering terms speed accuracy.