SVM-Based Feature Selection by Direct Objective Minimisation

Julia Neumann, Christoph Schnörr, Gabriele Steidl

Abstract

We propose various novel embedded approaches for (simultaneous) feature selection and classification within a general optimisation framework. In particular, we include linear and nonlinear SVMs. We apply difference of convex functions programming to solve our problems and present results for artificial and real-world data.
BibTeX
@inproceedings{neumann2004svm,
  author = {Julia Neumann and Christoph Schnörr and Gabriele Steidl},
  title = {{SVM-Based Feature Selection by Direct Objective Minimisation}},
  booktitle = {DAGM-Symposium},
  year = {2004},
  doi = {10.1007/978-3-540-28649-3\_26},
}

Loading related content...

Figure