NONLINEAR FEATURE EXTRACTION AND DIMENSION REDUCTION BY POLYGONAL PRINCIPAL CURVES
Author:
- 1 February 2006
- journal article
- research article
- Published by World Scientific Pub Co Pte Ltd in International Journal of Pattern Recognition and Artificial Intelligence
- Vol. 20 (01) , 63-78
Abstract
In this article we propose a polygonal principal curve based nonlinear feature extraction method, which achieves statistical redundancy elimination without loss of information and provides more robust nonlinear pattern identification for high-dimensional data. Recognizing the limitations of linear statistical methods, this article integrates local principal component analysis (PCA) with a polygonal line algorithm to approximate the complicated nonlinear data structure. Experimental results demonstrate that the proposed algorithm can be implemented to reduce the computation complexity for nonlinear feature extraction in multivariate cases.Keywords
This publication has 13 references indexed in Scilit:
- Learning and design of principal curvesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2000
- Application of neural network and FEM for metal forming processesInternational Journal of Machine Tools and Manufacture, 2000
- An Approach for Modeling Sheet Metal Forming for Process Controller DesignJournal of Manufacturing Science and Engineering, 1999
- Diagnosis of Multiple Fixture Faults in Panel AssemblyJournal of Manufacturing Science and Engineering, 1998
- Dimension Reduction by Local Principal Component AnalysisNeural Computation, 1997
- Self-Organization as an Iterative Kernel Smoothing ProcessNeural Computation, 1995
- Adaptive Principal SurfacesJournal of the American Statistical Association, 1994
- Principal curves revisitedStatistics and Computing, 1992
- Ice Floe Identification in Satellite Images Using Mathematical Morphology and Clustering about Principal CurvesJournal of the American Statistical Association, 1992
- Principal CurvesJournal of the American Statistical Association, 1989