• German
German

Main Navigation

Feature Selection for Rapidminer released

August 8, 2011 13:41


As part of project C1 - Feature selection in high dimensional data for risk prognosis in oncology - several new feature selection algorithms have been developed and publicly released. During his visit at the SFB, Viswanath Sivakumar implemented these algorithms as an extension to Rapidminer. The implementations are available for download on Sourceforge: RM-Featselext

  • Fast Correlation Based Filter (FCBF)
  • Shrunken Centroids – Prediction Analysis for Microarrays (PAM)
  • Backward Elimination via Hilbert-Schmidt Independence Criterion (BAHSIC)
  • Dense Relevance Attribute Group Selector (DRAGS)
  • Consensus Group Stable Feature Selector (CGS)



http://sourceforge.net/projects/rm-featselext/

SFB-876 NEWSLETTER
Newsletter RSS Twitter

NEWEST TECHREPORTS