An interface to calculate the "SeparationGain" for different eparation critiera used in various training algorithm There are two things: the Separation Index, and the Separation Gain Separation Index: Measure of the "purity" of a sample. If all elements (events) in the sample belong to the same class (e.g. signal or backgr), than the eparation index is 0 (meaning 100% purity (or 0% purity as it i ymmetric. The index becomes maximal, for perfectly mixed sample eg. purity=50% , N_signal = N_bkg Separation Gain: the measure of how the quality of separation of the sample increases by splitting the sample e.g. into a "left-node" and a "right-node" (N * Index_parent) - (N_left * Index_left) - (N_right * Index_right) this is then the quality crition which is optimized for when trying to increase the information in the system (making the best selection
| virtual | ~SeparationBase() |
| static TClass* | Class() |
| TString | GetName() |
| Double_t | GetSeparationGain(const Double_t& nSelS, const Double_t& nSelB, const Double_t& nTotS, const Double_t& nTotB) |
| virtual Double_t | GetSeparationIndex(const Double_t& s, const Double_t& b) |
| virtual TClass* | IsA() const |
| TMVA::SeparationBase& | operator=(const TMVA::SeparationBase&) |
| virtual void | ShowMembers(TMemberInspector& insp, char* parent) |
| virtual void | Streamer(TBuffer& b) |
| void | StreamerNVirtual(TBuffer& b) |
| TString | fName | name of the concrete Separation Index impementation |
Separation Gain: the measure of how the quality of separation of the sample increases by splitting the sample e.g. into a "left-node" and a "right-node" (N * Index_parent) - (N_left * Index_left) - (N_right * Index_right) this is then the quality crition which is optimized for when trying to increase the information in the system (making the best selection
Return the separation index (a measure for "purity" of the sample")