/2019-MLJ-Efficient-feature-selection-using-shrinkage-estimators

Matlab code for the methods presented in the paper "Efficient Feature Selection Using Shrinkage Estimators"

Primary LanguageMATLAB

2019-Efficient Feature Selection Using Shrinkage Estimators

Matlab code for the methods presented in the paper:
Efficient Feature Selection Using Shrinkage Estimators, K. Sechidis, L. Azzimonti, A. Pocock, G. Corani, J. Weatherall, G. Brown (to be appear in Machie Learing Journal)

Novel Shrinkage estimators (Section 3)

  • mi_Ind_JS.m - Implements our shrinkage estimator for mutual information (Section 3.2)
  • cmi_Ind_JS.m - Implements our shrinkage estimator for conditional mutual information (Section 3.3)

Novel High-order FS criteria (Section 4)

  • JMI3.m - Implements our FS method JMI-3 that takes into account third-order feature interactions (Section 4.2)
  • JMI4.m - Implements our FS method JMI-4 that takes into account fourth-order feature interactions (Section 4.2)
  • CMIM3.m - Implements our FS method CMIM-3 that takes into account third-order feature interactions (Section 4.2)
  • CMIM4.m - Implements our FS method CMIM-4 that takes into account fourth-order feature interactions (Section 4.2)

Tutorial

The tutorial 'Tutorial_Shrinakge_FS.m' presents how to select features using our suggested methods