Neural Computation, Vol 9, 1493-1516, Copyright © 1997 by The MIT Press

LETTERS

Dimension Reduction by Local Principal Component Analysis

Nandakishore Kambhatla and Todd K. Leen

Reducing or eliminating statistical redundancy between the components of high-dimensional vector data enables a lower-dimensional represen-tation without significant loss of information. Recognizing the limitations of principal component analysis (PCA), researchers in the statistics and neural network communities have developed nonlinear extensions of PCA. This article develops a local linear approach to dimension reduction that provides accurate representations and is fast to compute. We exercise the algorithms on speech and image data, and compare performance with PCA and with neural network implementations of nonlinear PCA. We find that both nonlinear techniques can provide more accurate representations than PCA and show that the local linear techniques outperform neural network implementations.

This article has been cited by other articles:

				G. E. Hinton and R. R. Salakhutdinov Reducing the dimensionality of data with neural networks. Science, July 28, 2006; 313(5786): 504 - 507. [Abstract] [Full Text] [PDF]

				E. Lopez-Rubio, J. M. Ortiz-de-Lazcano-Lobato, J. Munoz-Perez, and J. Antonio Gomez-Ruiz Principal Components Analysis Competitive Learning Neural Comput., November 1, 2004; 16(11): 2459 - 2481. [Abstract] [Full Text] [PDF]

				S. T. Roweis and L. K. Saul Nonlinear Dimensionality Reduction by Locally Linear Embedding Science, December 22, 2000; 290(5500): 2323 - 2326. [Abstract] [Full Text]