Forschung / Research

Theory of neural vector quantization

  • self-organizing maps
  • neural gas
  • (generalized) learning vector quantization
  • semi-supervised learning
  • functional principal component learning
  • fuzzy classification and clustering

Information theoretic learning

  • magnification and magnification control in neural maps
  • divergence based vector quantization
  • information optimum learning

Non-standard metrics in vector quantization

  • relevance learning
  • functional metrics and functional relevance learning
  • adaptive transformation invariant metrics
  • generalized metrics and dissimilarities for machine learning
  • kernel based metrics for online learning

Prototype-based classification learning vector quantization

  • adaptive classifcation learning by learning vector quantization (LVQ)
  • robust and secure classification by LVQ
  • LVQ classication based on statistical evaluation measures beyond accuracy (precision, specificity, F-measure etc.)
  • classification task dependent outlier detection
  • class representive and class border sensitive classification learning

Evolutionary algorithms

  • topologically structure island models
  • neighborhood cooperativeness for recombination

High-dimensional data analysis

  • hyperspectral remote sensing data analysis
  • mass spectrometry data analysis in medicine and biology
  • class visualization and projection