Prototypes and matrix relevance learning in complex fourier space

M. Straat, M. Kaden, M. Gay, T. Villmann, Alexander Lampe, U. Seiffert, M. Biehl, F. Melchert

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

12 Citations (Scopus)

Abstract

In this contribution, we consider the classification of time-series and similar functional data which can be represented in complex Fourier coefficient space. We apply versions of Learning Vector Quantization (LVQ) which are suitable for complex-valued data, based on the so-called Wirtinger calculus. It makes possible the formulation of gradient based update rules in the framework of cost-function based Generalized Matrix Relevance LVQ (GMLVQ). Alternatively, we consider the concatenation of real and imaginary parts of Fourier coefficients in a real-valued feature vector and the classification of time domain representations by means of conventional GMLVQ.
Original languageEnglish
Title of host publication12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)
PublisherIEEEXplore
Pages1-6
Number of pages6
ISBN (Electronic)978-1-5090-6638-4
DOIs
Publication statusPublished - 31-Aug-2017
Event12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM) - Nancy, France
Duration: 28-Jun-201730-Jun-2017

Conference

Conference12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)
CountryFrance
CityNancy
Period28/06/201730/06/2017

Keywords

  • Calculus
  • Discrete Fourier transforms
  • Frequency-domain analysis
  • Prototypes
  • Time series analysis
  • Time-domain analysis
  • Training
  • Classification
  • Learning Vector Quantization
  • dimensionality reduction
  • functional data
  • relevance learning
  • supervised learning

Cite this