Dimension Reduction by Local Principal Component Analysis
Nandakishore Kambhatla, Todd K. Leen · Neural Computation · 1997
Reducing or eliminating statistical redundancy between the components of high-dimensional vector data enables a lower-dimensional representation 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.
Cited by 717 works. doi:10.1162/neco.1997.9.7.1493
Authors
References
- Multilayer feedforward networks are universal approximators
- Numerical Recipes, The Art of Scientific Computing
- Multilayer feedforward networks are universal approximators
- Vector Quantization and Signal Compression
- The JPEG still picture compression standard
- On the approximate realization of continuous mappings by neural networks
- Vector Quantization
- Locally Weighted Regression: An Approach to Regression Analysis by Local Fitting
- Detection of signals by information theoretic criteria
- Simplified neuron model as a principal component analyzer