Limitations of nonlinear PCA as performed with generic neural networks

Authors:
E. C. Malthouse
Affiliations:
Northwestern Univ., Evanston, IL
Venue:
IEEE Transactions on Neural Networks
Year:
1998

Citing 0
Cited 13

A Unified Model for Probabilistic Principal Surfaces

IEEE Transactions on Pattern Analysis and Machine Intelligence
Estimating the Intrinsic Dimension of Data with a Fractal-Based Method

IEEE Transactions on Pattern Analysis and Machine Intelligence
Non-linear factor analysis (NLFA) with feedforward networks performs non-linear data reduction with extraction of linear scores

Second international workshop on Intelligent systems design and application
Data visualisation and manifold mapping using the ViSOM

Neural Networks - New developments in self-organizing maps
Complex-valued neural networks for nonlinear complex principal component analysis

Neural Networks
2007 Special Issue: Nonlinear principal component analysis of noisy data

Neural Networks
2008 Special Issue: On multidimensional scaling and the embedding of self-organising maps

Neural Networks
Nonlinear Principal Manifolds --- Adaptive Hybrid Learning Approaches

HAIS '08 Proceedings of the 3rd international workshop on Hybrid Artificial Intelligence Systems
Information Maximization in a Linear Manifold Topographic Map

Neural Processing Letters
Adaptive nonlinear manifolds and their applications to pattern recognition

Information Sciences: an International Journal
Model-based predictive control for spatially-distributed systems using dimensional reduction models

International Journal of Automation and Computing
An improved PCA fault detection for the diagnosis

ACMOS'07 Proceedings of the 9th WSEAS international conference on Automatic control, modelling and simulation
An algorithm that recognizes and reproduces distinct types of humanoid motion based on periodically-constrained nonlinear PCA

RoboCup 2004

Quantified Score

Hi-index	0.00

Visualization

Abstract

Kramer's (1991) nonlinear principal components analysis (NLPCA) neural networks are feedforward autoassociative networks with five layers. The third layer has fewer nodes than the input or output layers. This paper proposes a geometric interpretation for Kramer's method by showing that NLPCA fits a lower-dimensional curve or surface through the training data. The first three layers project observations onto the curve or surface giving scores. The last three layers define the curve or surface. The first three layers are a continuous function, which we show has several implications: NLPCA “projections” are suboptimal producing larger approximation error, NLPCA is unable to model curves and surfaces that intersect themselves, and NLPCA cannot parameterize curves with parameterizations having discontinuous jumps. We establish results on the identification of score values and discuss their implications on interpreting score values. We discuss the relationship between NLPCA and principal curves and surfaces, another nonlinear feature extraction method