a simple Binary search tree including volume search method
void | TMVA::BinaryTree::DeleteNode(TMVA::Node*) |
void | DestroyNode(TMVA::BinarySearchTreeNode*) |
void | Insert(TMVA::Event*, TMVA::Node*) |
Bool_t | InVolume(const vector<Float_t>&, TMVA::Volume*) const |
TMVA::BinarySearchTreeNode* | Search(TMVA::Event*, TMVA::Node*) const |
Double_t | SearchVolume(TMVA::Node*, TMVA::Volume*, Int_t, vector<const TMVA::BinarySearchTreeNode*>* events) |
TMVA::MsgLogger | TMVA::BinaryTree::fLogger | message loggera |
Int_t | TMVA::BinaryTree::fNNodes | total number of nodes in the tree (counted) |
UInt_t | fCurrentDepth | internal variable, counting the depth of the tree during insertion |
vector<Float_t> | fMax[2] | RMS for signal and background for each variable |
vector<Float_t> | fMeans[2] | mean for signal and background for each variable |
vector<Float_t> | fMin[2] | RMS for signal and background for each variable |
Double_t | fNEventsW[2] | Number of events per class, taking into account event weights |
UInt_t | fPeriod | periode (number of event variables) |
vector<Float_t> | fRMS[2] | RMS for signal and background for each variable |
Bool_t | fStatisticsIsValid | flag if last stat calculation is still valid, set to false if new node is insert |
vector<Double_t> | fSum[2] | Sum for signal and background for each variable |
Double_t | fSumOfWeights | Total number of events (weigthed) counted during filling |
vector<Double_t> | fSumSq[2] | Squared Sum for signal and background for each variable |
copy constructor that creates a true copy, i.e. a completely independent tree
insert a new "event" in the binary tree set "eventOwnershipt" to kTRUE if the event should be owned (deleted) by the tree
private internal function to insert a event (node) at the proper position
search the tree to find the node matching "event"
Private, recursive, function for searching.
create the search tree from the events in the DataSet NOTE: the tree given as argument MUST NOT contain transformed variables, otherwise double transformation would occur
create the search tree from the event collection
using ONLY the variables specified in "theVars"
search the whole tree and add up all weigths of events that lie within the given voluem
recursively walk through the daughter nodes and add up all weigths of events that lie within the given volume
test if the data points are in the given volume
set the periode (number of variables) return periode (number of variables)
{ fPeriod = p; }
access to mean for signal and background for each variable
{ return fMeans[sb==Types::kSignal?0:1][var]; }
access to RMS for signal and background for each variable
{ return fRMS[sb==Types::kSignal?0:1][var]; }
access to Minimum for signal and background for each variable
{ return fMin[sb==Types::kSignal?0:1][var]; }
access to Maximum for signal and background for each variable
{ return fMax[sb==Types::kSignal?0:1][var]; }
recursive search through daughter nodes in weight counting