Hierarchical feature maps for non-linear component analysis

M. Herrmann, R. Der, G. Balzuweit

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Based on earlier work on self-organizing maps with adaptive local neighborhood widths suitable for construction of principal manifolds, we propose an algorithm for hierarchical maps of heterogeneous high-dimensional data onto a structurally similar output space. Instead of a fixed output grid a network structure evolves that is locally orthogonal, but globally shaped by prominent data features. These features form principal manifolds in subspaces being determined by earlier hierarchical levels. The algorithm allows for an efficient separation of the interdependent learning tasks of acquiring optimal maps, learning parameters, and network structure
Original languageEnglish
Title of host publicationNeural Networks, 1996., IEEE International Conference on
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1390-1394
Number of pages5
Volume2
ISBN (Print)0-7803-3210-5
DOIs
Publication statusPublished - 1 Jun 1996

Keywords

  • learning (artificial intelligence)
  • self-organising feature maps
  • adaptive local neighborhood widths
  • heterogeneous high-dimensional data
  • hierarchical feature maps
  • hierarchical levels
  • locally orthogonal network structure
  • network structure
  • nonlinear component analysis
  • optimal maps acquisition
  • parameter learning
  • principal manifolds
  • prominent data features
  • self-organizing maps
  • Convergence
  • Gaussian processes
  • Information representation
  • Network topology
  • Neural networks
  • Neurons
  • Proposals
  • Self organizing feature maps

Fingerprint Dive into the research topics of 'Hierarchical feature maps for non-linear component analysis'. Together they form a unique fingerprint.

Cite this