Applying a trained ML model to unlabeled data

TMVA
, ,
1

Hello all,

  • I’m doing a binary classification between signal (new resonance) and background (SM) using MLP of the TMVA.
  • I’m using already weighted TTree events for training the model.
  • After doing the training, I apply the trained model read from the .xml file to unlabeled data in a weighted TTree, but I get un-weighted score (DNN response ) at the end, what should I do to get the correct weighted score?

Thanks in advance
Magdy

2

Hi,

I am not sure I have understood your problem. When you apply your trained model to your events you will get a DNN response that should be weighted by the weight of that particular event. You don’t need to pass the weight as input to the DNN.

Lorenzo