TMVA::evaluate function for BDT score of each event

Hii all,

I am wondering, how the evaluation is really being done when I am applying the TMVA weights (obtained from a BDT training) to classify the events in a test sample as signal like or background like. How the BDT score is obtained for each event ??

Can I have a mathematical formulation of this evaluation method of BDT ??

How this BDT score is calculated for each event from the votes given by the boosted decision trees during evaluation ???

By summing the responses of all trees with the appropriate weights. The response of each tree is calculated by following all decision branches of the tree using the observables of each event.

The formulae for the evaluation can be found in the TMVA User’s guide:

The BDT score is simply the weighted sum of the responses for all trees, possibly with a transformation to put it in a specific interval such as [0, 1]